Tägliche Publikationen

General-purpose language models can reason and synthesize knowledge, but complex work also requires sustained interaction with files, information sources, and executable code, together with state maintenance, failure recovery, and verifiable delivery. We call this working capability: sustained, verifiable progress toward a real-world objective. Apodex 1.1 develops this capability along two complementary dimensions. Environment Scaling expands the diversity and verifiability of executable file, search, and code environments, while Agentic Coordination Scaling trains agents to decompose long-horizon tasks, delegate parallel work, integrate asynchronous results, and replan. A shared execution harness and AgentOS maintain task state and provenance across tools and agents, and training turns environment trajectories and coordination traces into reliable behavior. Across complex professional work, finance, scientific research, mathematics, coding, and search, Apodex 1.1 reaches the leading performance band despite using a substantially smaller model than many frontier systems. The 35B-parameter Apodex 1.1 Mini further retains strong working capability in a locally deployable form. These results ground agentic intelligence in useful, verifiable work completed over time and advance our goal of building a Heavy-Duty Solver for ambitious, long-running tasks.

We present EchoWM, an omnimodal world model for enterable generative media that responds to continuous navigation while jointly generating 720p video, environmental sound, music and speech. We organize interaction around camera intent: in first-person scenes, it specifies observer motion, while in third-person scenes, camera--character dynamics are learned from data without view-specific controllers. Discrete commands and continuous poses are mapped to a shared metric-scale relative 6-DoF trajectory, with dataset-level calibration preserving motion magnitude across heterogeneous data. To jointly learn audio-visual generation and trajectory control, we construct a complementary data engine and adopt progressive training followed by autoregressive post-training for long-horizon generation. Extensive evaluations show that \model achieves strong trajectory following and high visual quality on public world-model benchmarks, supporting both first- and third-person interaction across varied subjects, and maintaining synchronized environmental sound and speech over long-horizon generation.

E-commerce live streaming requires omni-modal understanding of noisy, temporally extended streams, where product facts are distributed across speech, video frames, product images, overlaid text, and user queries. We present TLive-Omni, an omni-modal understanding model tailored to live-commerce scenarios. It maps image, video, audio, and text inputs into a unified representation space. For long-form live streaming analysis, we introduce Per-vGrid, a timestamped token organization that groups each video grid with its temporally corresponding audio within explicit boundary tokens to facilitate temporal alignment. We design a three-stage supervised training recipe that progressively develops live-commerce understanding, from omni-modal perception to instruction-following responses. We then propose Faithful-RFT, a reinforcement fine-tuning stage that further improves answer faithfulness and expression quality while meeting real-time demands, scoring final responses directly with task-verifiable feedback rather than optimizing for reasoning-style exploration during rollout. Moreover, TLive-Omni is supported by a scenario-oriented atomic capability taxonomy and a compact data production engine that converts live-commerce audio, image, and video streams into training signals for speech recognition, speaker analysis, product visual grounding, text recognition, temporal grounding, video dense caption, and omni-modal QA, etc. For scalable training, a synchronized length-grouped sampler reduces padding while preserving comparable workloads across workers, while a lightweight dynamic sampling strategy regenerates rollout groups with near-zero reward variance to maintain meaningful relative advantages for GRPO. Experiments on e-commerce live streaming benchmarks demonstrate strong performance across live-commerce domain tasks, together with excellent generalization on general benchmarks.

Existing image editing frameworks predominantly follow the training paradigm of text-to-image diffusion models. However, extending this paradigm to image editing highlights two inherent discrepancies, specifically, the insufficient attention to edit concept granularity and the training inefficiency caused by sparse supervision signals. To address these issues, we establish a comprehensive hierarchical taxonomy featuring over 1,000 fine-grained edit concepts and build ConceptEdit-12M, a massive dataset of 12 million high-quality editing pairs via an improved synthesis framework. This library-driven approach effectively rectifies the distribution collapse of generated data while ensuring high data fidelity. Furthermore, we propose a dense supervision training strategy that synthesizes multiple non-interfering concepts into single image pairs. By providing richer learning signals, this strategy significantly enhances both training efficiency and overall model performance. Training results validate our strategy, significantly outperforming prior works. Finally, we present ConceptEdit-Bench, a granular evaluation suite designed to diagnose model capabilities across a vast array of real-world scenarios.

As on-device LLM agents evolve into personal copilots, the mobile operating system has become a key testbed for this paradigm, making rigorous capability evaluation essential. Yet existing benchmarks fall into two camps, each with a critical blind spot: GUI-centric benchmarks test surface-level screen manipulation while overlooking background tool use and long-horizon planning, whereas static function-calling benchmarks rely on offline API matching that is detached from real runtime constraints. To close this gap, we present MobilePA-Bench, an interactive, stateful, and tool-centric benchmark for evaluating the tool-calling and planning abilities of mobile planning agents. MobilePA-Bench runs on an executable sandbox that maintains live application databases and returns structured feedback, spanning 13 functional domains and 212 realistic mobile tools. Beyond basic tool use, it evaluates a central planning agent along three advanced dimensions: (1)~Sub-agent Collaboration---decomposing a complex task and delegating specialized work to capable sub-agents; (2)~Memory Usage---recalling stored memories, user profiles, and past preferences to resolve implicit requests; and (3)~Skill Usage---invoking pre-packaged composite skills instead of planning every step from scratch. Extensive experiments show that current frontier LLMs remain unreliable in mobile settings: performance drops sharply under strict tool ordering, permission limits, and unexpected runtime errors. By pairing an interactive function-calling sandbox with evidence-based verification, MobilePA-Bench serves as both a practical diagnostic benchmark and an interactive foundation for agentic reinforcement learning---accelerating the development of dependable mobile agents.

Language models are sequential processors, but long-horizon agency requires external information and computation beyond model weights and active context. Prime Agent is an open-source harness for long-horizon evaluation and coding-agent workflows. A persistent IPython REPL follows the Recursive Language Model abstraction for programmatic context processing and test-time compute, while Continual Harness preserves histories, memories, skills, prompts, and subagent specifications across trajectories. Recursive subagents coordinate through direct agent-to-agent communication, and the Agents View lets humans inspect and manage daemon-backed sessions. Prime Agent standardizes execution, recovery, verification, and resource accounting while leaving strategy construction to the model. This low-friction, expressive membrane prevents harness failures from becoming model failures and pushes measurement toward the model's true maximal underlying capability. Prime Agent raises ARC-AGI-3 RHAE Best@1 from 30% to 95.5% and matches or exceeds native and popular harnesses across long-context coding, GPU-kernel generation, emulator construction, and autonomous nanoGPT speedruns. On Factorio, we find refinement allows for continuous technology progression and dedicated subagents enable parallelized work. Code is available at https://github.com/PrimeIntellect-ai/prime-agent.

While text-to-3D generation has advanced rapidly, achieving high geometric fidelity at low inference cost remains challenging. Existing text-to-3D methods either decode discrete shape tokens autoregressively or iteratively refine global 3D representations with diffusion or flow models. However, autoregressive decoding is sequential and cannot revise errors, whereas diffusion and flow-matching models repeatedly process the full representation, making high-quality generation increasingly expensive. In this paper, we propose Block3D, a block-wise diffusion framework that partitions the discrete shape-token sequence into contiguous blocks, generates the blocks autoregressively, and jointly denoises all tokens within the current block. To alleviate error accumulation, we introduce confidence-guided intra-block correction, which revises low-confidence tokens before each block is finalized. On a held-out set from TRELLIS-500K, Block3D reduces mean end-to-end generation time from 25.71 seconds to 4.99 seconds, achieving a 5.15times speedup over the fine-tuned autoregressive baseline without sacrificing geometric fidelity.

World Action Models (WAMs) improve planning by incorporating future world evolution into action generation, yet existing methods allocate a fixed imagination budget to every scene. We propose RISE (Refining Imagination through SElective Rollout), a system-level adaptive imagination framework that makes sequential Roll/Stop decisions according to the expected planning benefit of continued rollout. At each step, a Latent Evaluator estimates the risk revealed by the current prefix and how much planning could improve if imagination continues, while a Rollout Gate weighs this expected benefit against additional computation cost. Since factual driving logs expose only one realized future, we further construct CounterDrive, a counterfactual dataset with diverse outcomes and risk levels, to enrich future dynamics and provide localized risk supervision. Each retained sample undergoes expert verification and annotation of trajectory validity, incident onset, and causal category, providing a reusable resource for safety-critical world-modeling research. Experiments on NAVSIM and nuScenes show that RISE achieves the best overall planning performance while reducing unnecessary rollout, with additional transfer results supporting its plug-in generality across WAM architectures.

User representation learning in real-world industrial scenarios is commonly scaled by increasing user amount, behavioral sequence length and model size. However, existing methods face two challenges: (i) Bottleneck for raw data scaling at billion-scale capacity, as performance exhibit diminishing performance gains with larger-scale raw text user behavioral input, which can be mitigated by tokenization. (ii) Lack of quantitative analysis of how tokenization configurations should scale with data size. In this report, we propose User Behavioral Densing Law for characterizing the quantitative relationship between data scale and the minimum sufficient tokenization capacity. Firstly, we conduct a pilot study on raw & tokenized scaling comparison on billion-scale Alipay dataset, revealing the raw data scaling bottleneck and the sustained gains enabled by tokenization. To derive the scaling pattern governing the minimum sufficient tokenization configuration at different data scales, theoretical analysis and systematic experiments are employed to summarize the quantitative scaling pattern. We find an approximately linear relationship between the logarithms of minimum sufficient tokenization capacity and input data size measured by tokens, and the scaling slope varies systematically with the tokenization method and data source, reflecting differences in representation-space redundancy and intra-source uniqueness. Guided by the proposed law, we further develop ALGN, an adaptive variable-length tokenization method that improves capacity allocation. Extensive experiments across diverse data sources, tokenization methods, and downstream tasks demonstrate the generalizability and reliability of the User Behavioral Densing Law, providing practical guidance for tokenization configuration selection in large-scale user representation learning. Moreover, ALGN outperforms existing baselines.

An interactive world model must follow the user's actions, remember the places it has shown, and stream in real time. The tension is structural: control wants a short horizon, memory wants an unbounded one. ReWorld separates the two during training and bounds them at inference. Mixed per-head attention windows confine most heads to the recent past while a small set of global heads attends over the entire history, and random head routing keeps either capability from binding to particular heads; random chunk dropping makes sparse histories in-distribution. At inference the whole past lives under a fixed budget: a bounded KV cache backed by a pose-indexed landmark bank, from which the model retrieves the landmarks nearest the current pose. A metric-scale-aligned data engine places eight sources -- Unreal-rendered fly-throughs, game roaming, and real-world footage -- on one physical action scale, so the same key press moves the camera the same distance in every source, and palindrome trajectories supply the revisit evidence that memory training needs. Distribution-matching distillation confined to a LoRA adapter then compresses sampling to four steps: one backbone serves both a high-fidelity multi-step mode and a real-time interactive one, streaming 704x1280 video across photorealistic, game-style, and stylized worlds. Under a three-axis protocol covering action following, long-horizon recall, and video quality, against six recent interactive world models it attains the best control fidelity (11.95^circ rotation error and the best camera-motion consistency) and the best generation quality; and on minute-long out-and-back rollouts (64\,s, 384 latents), its fixed 12-chunk cache still regenerates the starting view -- at rollout lengths where a sliding window has long evicted the evidence and full-KV attention runs out of memory.

Policy optimization (PO) for Large Language Models faces a stability--exploration trade-off, currently mediated by an action-side Policy-KL regularizer. This puts practitioners in a double bind: keeping Policy-KL constrains response behavior and consumes the action-side exploration budget, while dropping it leaves the optimization without an explicit drift control. We argue for an alternative that breaks the dilemma by moving regularization to the input side. As training progresses, the distribution over training queries induced by the current policy drifts unchecked from its pre-RL reference distribution. Concretely, Environment-Regularized Policy Optimization (ERPO) introduces a Query-KL (QKL) term that bounds this query distribution shift, together with a dataset-static reference-derived per-query weight that biases each per-query update toward queries typical under the reference. The QKL gradient flows strictly through the query likelihood; the response score function used by policy-gradient estimators does not appear in the QKL term, so QKL exerts no direct gradient pressure on the response distribution---exploration is preserved. ERPO plugs into GRPO/PPO/REINFORCE-style pipelines without additional forward passes. On six mathematical reasoning benchmarks, ERPO replaces the standard Policy-KL regularizer while achieving effective control over query distribution drift, delivering stronger accuracy and substantially more stable behavior under high-temperature decoding and long-horizon training.Our source code are available at https://github.com/alibaba/ERPO

Open-ended real-world interaction admits multiple valid behaviors: an agent may answer directly, ask for clarification, provide progress updates, or confirm before acting. This flexibility breaks a core assumption behind group-based RL: rollouts compared within a group are no longer guaranteed to be behaviorally comparable. As a result, reward-model preferences over interaction style can distort relative advantages and steer optimization toward reward-preferred behaviors rather than context-appropriate ones. We formalize this as a reward fairness problem and propose ARC (Advantage Regularization via Conditioning), a training recipe that restores fairer relative comparison through strategy-conditioned rollout grouping, together with hybrid rewards and entropy regularization. We study ARC in our proposed \inter, a novel paradigm for responsive, steerable, and execution-aware user-agent interaction that decouples user-visible communication from latent reasoning and tool use. \inter\ also provides the annotation and distillation pipeline for constructing \inter-86K, our strategy-annotated training corpus for supervised and RL training. Empirically, ARC substantially strengthens the core τ/τ^2 tool-use benchmarks, while \inter\ reduces time-to-first-token from 4.91s to 1.27s relative to a think-style baseline. Together, these results suggest that a central bottleneck in open-ended interactive learning is not only how agents are rewarded, but whether their behaviors are compared fairly in the first place. The ARC implementation and \inter-86K training data will be released.

Recent large language models (LLMs) can operate as coding agents that build complete games from natural language requests. Game development is especially demanding because program logic, visual and audio content, interfaces, interaction and playability must function together in one executable artifact. Measuring this capability therefore requires evaluation of both game product and the development process. Existing benchmarks often assess the game development capabilities of LLMs by evaluating the final artifact or an isolated development stage. Our analysis of complete human-agent development trajectories identifies three stages that together span the lifecycle of game development with a coding agent: initial game generation, bug diagnosis and repair, and optimization over multiple turns. Therefore, we introduce GameXpert-Bench, which operationalizes the three lifecycle stages as three complementary benchmark tracks. GameGen evaluates complete game creation from a single request in an empty workspace. GameFix evaluates diagnosis and repair when defects are reported or left for the agent to discover. GameOpt evaluates cumulative optimization through request chains seeded by real development trajectories between users and agents. We evaluate each track using live game interaction, deterministic behavioral tests, or final product criteria with regression checks. The suite contains 97 generation tasks across 11 genres; 100 repair tasks from 50 game levels verified by humans, each with 19-27 injected bugs; and 17 optimization chains with six turns and 102 requests. Across the three tracks, current agents are more reliable at producing playable foundations and implementing explicit requirements than at discovering defects, verifying runtime behavior, and preserving functionality across changes.

Recent agent benchmarks increasingly ground evaluation in executable environments, from code repair to web navigation, app APIs, and function calling. Yet completing consequential work beyond code requires more than producing a plausible response or valid tool call: agents must gather missing information over multiple turns, follow domain policies, coordinate dependent tools, and realize the correct persistent state transition without collateral effects. In this paper, we introduce Thinkingbox, a sandbox for tool-agent-user interaction that provides isolated MCP-compatible tool sessions, complete execution traces, and outcome evaluation over terminal backend state. Built on this sandbox, Thinkingbox-bench contains 507 policy-conditioned workflows across numerous scenarios, including retail, hospitality, auto insurance, neobank internal IT, and consulting IT/HR support. Each attempt is evaluated by task-specific executable checks that accept valid trajectories while rejecting wrong, missing, or extra effects; designated tasks additionally check required properties of the final response. Across proprietary and open-weight models, the strongest achieves 65.36% pass@1, but only 25.25% pass^20. Moreover, many failed trials show clean termination and valid state-changing actions, showing that response or tool-call-level signals are not clear proxies for end-to-end task completion. Thinkingbox-bench reveals a large gap between occasionally finding a successful trajectory and reliably completing stateful business tasks. We release both Thinkingbox and Thinkingbox-Bench: https://github.com/microsoft/thinkingbox

Serving large language models cheaply increasingly means shipping models that are both structurally compressed to a fraction of their parameters and quantized to 4 bits. Together these steps degrade reasoning, mathematics, coding, and long-context behavior enough to require a recovery, or healing, stage before deployment. The default recipe, quantization-aware training (QAT), re-fits the compressed, quantized model to hard labels; in our pipeline it converged slowly and collapsed past its peak. We adopted Quantization-Aware Healing (QAH) instead. Because a structurally compressed model is never independently trained at full precision, its bfloat16 checkpoint is a distillation-recovered approximation of the original; QAH distills the 4-bit student directly from the original, uncompressed model. On a GPT-OSS 120B to 60B to MXFP4 pipeline, the QAH student matches or beats its bfloat16 source on 7 of 9 benchmarks at roughly 4 times less weight memory and half the teacher's parameter count, and is released open-weight as Hypernova-60B. Against a matched QAT baseline it reaches a comparable peak about 7 times faster and stays stable under continued training, without hand-tuned early stopping. We also report deployment lessons, including a large, reproducible quality gap between distributed-training backends. Our aim is a recipe deployable without a multi-week hyper-parameter search.

Autonomous research systems are increasingly capable of executing long research workflows, yet automation alone does not ensure that the resulting process remains scientifically grounded. We introduce AutoResearch, a two-stage system that connects Idea Generation with Idea Execution to address both how research ideas are formed and how they are reliably established through experimentation. In Idea Generation, AutoResearch continuously integrates emerging research signals with accumulated domain knowledge, identifies transferable mechanistic insights, and uses multi-model generation and cross-review to produce grounded, testable research plans. In Idea Execution, coordinated agents decompose these plans into experiments, iteratively implement and diagnose them, and employ independent evidence-based review before accepting research conclusions. Across representative settings in cross-modal retrieval, systems optimization, and benchmark-driven machine learning, AutoResearch turns generated ideas into measurable progress, detects and corrects unreliable experimental results, and makes evidence-conditioned decisions to continue, revise, or terminate research directions. For example, on RSICD benchmark, an AutoResearch-generated idea improves mean Recall from 32.84 to 34.69, while recording only 5 audit-confirmed issue events compared with 11-27 for other autonomous research systems. These results demonstrate a research process in which meaningful insight is grounded before experimentation and conclusions are grounded before acceptance: Insight In, Hallucination Out.

On-policy distillation (OPD) has emerged as an effective framework for post-training language models by pairing student-generated trajectories with dense token-level supervision from a teacher. However, OPD implicitly assumes that teacher-derived rewards are an appropriate proxy for reasoning progress, and therefore treats all teacher feedback equally during policy optimization. While in practice, this assumption does not always hold. We observe that teacher-derived rewards often conflict with genuine reasoning progress, as reasoning steps with clear reasoning advancement may still receive lower distillation rewards, simply due to deviation from teacher's outputs. To address this mismatch, we propose Reasoning-Progress-Aware Reward Filtering for On-Policy Distillation (R2-OPD), which constructs two within-trajectory rankings of reasoning spans, one from teacher-derived rewards and the other from independently estimated progress reward. Distillation rewards are selectively suppressed whenever the two rankings disagree, reducing supervision that conflicts with reasoning progress while preserving effective teacher guidance. Our approach shows consistent improvement over standard OPD especially regarding reasoning performances.

We present a novel approach to efficient LLM harness optimization through adaptive validation task selection. Harness optimization iteratively rewrites the harness code based on validation performance, enabling substantial performance gains without updating the underlying model weights. Existing approaches, however, evaluate a fixed validation set in full at every iteration, incurring substantial evaluation costs even on tasks that become less discriminative as the harness evolves. We propose Task-CoEvolve, which co-evolves the validation tasks with the harness by addressing two challenges: selecting informative tasks and estimating full-set performance from partial evaluations. Task-CoEvolve builds on the observation that tasks on which candidate harnesses disagree are more informative for distinguishing among them than tasks that are consistently solved or failed. It uses variance-weighted sampling based on past outcomes to focus evaluation on tasks near the capability frontier, with the sampling distribution adapting as the harness evolves. It then estimates full-set scores from the sampled tasks by accounting for their sampling probabilities, enabling consistent comparisons across iterations despite evaluating different subsets. Experiments on online text classification and Terminal-Bench 2.1 show that Task-CoEvolve consistently outperforms subset-based baselines and matches the final performance of full-set search while reducing the number of evaluations during optimization by 80%. Code will be released at https://github.com/Agent4Science-UTokyo/Task-CoEvolve.

Large language models are increasingly expected to execute complex workflows whose success depends on maintaining interdependent constraints and producing artifacts that satisfy strict end-to-end verification. Yet successful execution experience is typically lost after a single run, forcing subsequent models to rediscover strategies and failure modes from scratch. We study whether such experience can instead be externalized and reused through EvoMap, where verifier-confirmed execution trajectories are consolidated into structured Gene. To evaluate this setting, we introduce the Long-Workflow Benchmark (LongWoF-Bench), comprising 778 machine-verifiable tasks across code generation, agent-environment synthesis, mathematical reasoning, and rule following. On the 252 tasks with verifier-confirmed Opus trajectories, evolved EvoMap Gene outperform Skill across all seven evaluated models by 8.7-15.5 percentage points, with the gains extending to consumer models from different model families. In contrast, reference-distilled Gene do not exhibit the same advantage, indicating that compact representation alone is insufficient and that Gene utility is closely associated with verified experience provenance. For Claude Opus, Gene reuse also completes 39 more tasks than Skill while reducing solve-time token consumption by 9.9%. Together, these results show that verified execution experience can be retained and shared as a reusable external resource, enabling models to improve long-workflow completion without repeatedly paying the full cost of experience discovery.

Speech-based applications pass spoken queries through automatic speech recognition (ASR) before any retrieval module, so ASR errors enter the pipeline as a fixed upstream constraint. We empirically test whether two extensions to standard retrieval-augmented generation (RAG), entity-graph linking and iterative reformulation, absorb or amplify these errors. Using four English accents synthesized through neural TTS, we evaluate four RAG configurations on three multi-hop QA benchmarks (HotpotQA, 2WikiMultiHopQA and MuSiQue) against a clean-text oracle. Although the structurally richer configurations generally retain higher absolute F1 under ASR input, both extensions amplify the error: the F1 gap from clean text to the highest-WER accent is 36-67% larger under their combination than under naive dense retrieval, on all three benchmarks. The dominant failure mode is corruption of one or more query entities, accounting for 87-96% of degradation cases on 2WikiMultiHopQA across all four methods. Two lightweight surface-form mitigations leave most of the gap intact, indicating that downstream retrieval structure amplifies remaining entity errors. We release code and data at https://github.com/ZhenghuaBao/spoken-multihop-rag .

Long-context prefill in large language models (LLMs) incurs substantial computation and memory traffic because dense self-attention computes quadratic query-key scores. Existing methods either use a uniform low-precision path or select token interactions, leaving spatial precision routing over hardware-aligned score tiles outside fused dense attention. We introduce TileMix, a tile-centric precision-routing kernel that makes numerical precision an executable spatial decision over score-tile groups within fused dense attention. TileMix partitions the attention matrix into hardware-aligned score tiles, packs routing decisions into compact bitmasks, and dispatches each tile group through FP16 or INT8 score computation while both paths update a shared online-softmax state. Scalable precision grouping lets each routing bit govern multiple adjacent key tiles, preserving hardware-aligned compute tiles and compact metadata at long contexts. By routing all legal tile groups, TileMix preserves dense token connectivity, requires no training, and supports grouped-query attention, variable-length batches, and INT8 key/value caches. Across LongEval, LV-Eval, and A100 prefill benchmarks on LLaMA, Qwen, and Vicuna, TileMix recovers long-context quality lost under uniform INT8 and improves prefill throughput over FP16, yielding a controllable accuracy-efficiency frontier across model families. The implementation is available at https://github.com/HanzhiZhang-Ulrica/TileMix.

Industrial technical reports contain high-value knowledge for maintenance, troubleshooting, and product engineering, but their heterogeneous structure (dense prose, specifications, tables) makes them difficult to index and reason over with standard retrieval and QA pipelines, and no public instruction-tuning or benchmark datasets are built from such documents. We address this gap with Industrial-Instruction, contributing (i) two open QA datasets built from real industrial technical reports and (ii) the end-to-end pipeline that produces them. Using 906 public Panasonic documents (7,525 pages), we apply layout-aware extraction, build a semantic retrieval index, and synthesize multiple-choice QA grounded in retrieved evidence under five query-document relationships (irrelevant retrieval, single-/multi-document support, single-/multi-document answer). After filtering an initial 23.9k generated samples, each dataset provides approximately 13.6k QA pairs with source documents and a held-out benchmark split. Fine-tuning small open LLMs (under 10B parameters) improves Set-Match Accuracy from 28.5% to 42.0% and F1 from 46.6% to 63.5% on the Panasonic benchmark. We release two parallel versions built by the same pipeline: one generated with the open-weight Qwen3-30B-A3B-Instruct model and one with the closed, API-based Claude-Opus-4.6 model, enabling a direct comparison of open- versus frontier-model data generation. The Claude-Opus-4.6 dataset yields a cleaner raw corpus and larger fine-tuning gains, at roughly two orders of magnitude higher cost. MMLU evaluation shows models trained on the Claude-Opus-4.6 data retain essentially all general knowledge, versus a small but measurable forgetting effect for the Qwen-generated data. Together, these datasets and pipeline offer a practical, reproducible path toward scalable industrial benchmarks and training data from real-world documentation.

Search-augmented LLMs increasingly mediate everyday consumer recommendations by retrieving live web content. This creates a new risk: LLM recommenders may consume web content that Generative Engine Optimization (GEO) operators have polluted to mislead them. We ask: to what extent do they become unwitting promoters of fake products? We introduce FORGE (Fake Online Recommendations in Generative Environments), which locally rewrites real products in a frozen set of retrieved web pages into fake ones and measures how often the LLM recommends the fake product, across 225 real products in 15 categories and 5 consumer scenarios. Across 12 commercial and open-weights LLMs, all models are vulnerable: a single polluted page yields fooled rates of up to 27%, while the full top-3 replacement raises this to 73.8%. Vulnerability varies across categories, increasing when models lack stable prior knowledge of the products. Reasoning does not mitigate this vulnerability; instead, it often generates spurious social proof to justify false recommendations. None of the four defenses is adequate: the skepticism prompt can exacerbate vulnerability much like reasoning, the two consensus filters risk suppressing legitimate products, and credibility re-ranking helps every model but removes only a sixth of the fakes. We release the FORGE benchmark and the evaluation code at https://github.com/leoluolol/forge-benchmark.

As Retrieval-Augmented Generation (RAG) shifts toward diverse portfolio generation, it is stymied by two critical bottlenecks: flawed measurement of evidence utilization, and suboptimal context budget allocation. We resolve both sequentially. To resolve measurement, we expose a pervasive ``diagnostic illusion'': standard relevance proxies fail catastrophically on hard negatives. We replace them with an efficient causal leave-one-out probe that accurately isolates generative reliance and formally calibrates the structural dilution of LLM attention. To resolve allocation, we deploy this causal probe in a deconfounded factorial grid. We prove that the prevailing strategy of monolithic context widening is an architectural trap penalized by relevance decay. Instead, allocating compute iteratively across multiple sequential generations drives transformative portfolio recall gains of 16.7--20.5 absolute percentage points, scaling robustly up to 32B models. Finally, we unify these solutions into a deployable closed-loop submodular scheduler. Augmented by an attribution-steered contrastive decoder to override LLM attention inertia, our architecture systematically forces fresh evidence integration. By dominating classical open-loop baselines, we establish sequential, feedback-driven orchestration as the definitive paradigm for generative search. Our code, data, and causal measurement instruments are available at https://github.com/PeiYangLiu/ascp.

A retrieval-augmented QA system can return different answers after an index expansion even when its requested model identifier, prompt, retrieval policy, evidence depth, rendering, and exposed generation controls are held fixed. Aggregate accuracy may hide these changes when gains and losses cancel, while ordinary generation variability makes one-shot comparisons overstate update effects. We call the hidden phenomenon accuracy-blind answer churn and introduce the Snapshot Compatibility Audit, which estimates excess answer churn by subtracting same-snapshot repeat disagreement from cross-snapshot disagreement. We instantiate it by expanding one frozen FineWeb prefix from one to seven shards. In a preregistered 400-question Natural Questions study, normalized-exact and blinded-semantic excess churn are 6.44 and 10.25 percentage points while exact-match accuracy changes by only -1.50 points. A post-hoc analysis finds repeat-stable semantic flips on 40/400 questions. A separately preregistered 200-question TriviaQA study yields smaller, directionally consistent excess churn while exact-match accuracy moves in the opposite direction. An outcome-blind post-hoc 100-question subset replication with a second DeepSeek generator and serving configuration finds 8.75 pp of semantic excess churn even as exact match rises by 3.00 percentage points. Answer-level compatibility can therefore fail without a conspicuous or consistently directed utility shift. Retrieval-augmented releases should audit compatibility alongside utility.

With the rapid progress of diffusion models and large-scale video generation, generative world models are increasingly expected to replace traditional simulators, including physics engines, game engines, and reinforcement-learning environments. Yet the remaining distance from generation to simulation lacks a systematic assessment. We present a capability-based study using an external yardstick: eight capabilities of a traditional simulator, namely asset construction, physics engine, interaction, controllability, stability, state feedback, diversity, and evaluation metrics. We trace three main technical routes--latent dynamics, video generation, and joint-embedding prediction--and map exactly 200 representative works published from 2018 to June 2026 onto these capabilities. Our analysis shows that world models have achieved functional substitution in interaction and controllability for specific scenarios, but remain short of traditional simulators in formal guarantees of physical laws, structured state feedback, and reproducible long-horizon evolution. State feedback is the most neglected cross-route shortcoming: only 6 of 163 implementation papers expose a runtime interface for querying entity states or physical parameters. We identify six research directions: formalized physics, a unified action interface, first-class state feedback, long-horizon stability, downstream-utility evaluation, and cross-route hybridization. Project page: https://github.com/AtongWang/world-model-simulators

Full-length RNAs, particularly messenger RNAs, often exceed the context lengths used to pretrain existing RNA foundation models, limiting complete-transcript modeling at single-nucleotide resolution. We present RIBOSPAN, a 1.61-billion-parameter bidirectional RNA foundation model natively pretrained with context lengths up to 10,240 nt. RIBOSPAN combines dense bidirectional self-attention, single-nucleotide tokenization, and attention-isolated sequence packing to enable high-resolution modeling of complete long RNAs. We evaluate the model through nucleotide reconstruction, a controlled long-context representation benchmark, and frozen RNA-type representation analysis. Native 10K pretraining preserves strong reconstruction at 10,240 tokens, while continued pretraining with 40% masking improves recovery under heavy corruption while preserving representation quality. The long-context benchmark further shows that native 10K models maintain strong contextual responsiveness and context-specific representation separation while keeping perturbation-induced representation changes highly localized. Inference-time YaRN scaling recovers much of the contextual organization lost by direct extrapolation of short-context models, but induces substantially greater distal representation diffusion. Frozen-representation evaluations further demonstrate state-of-the-art RNA representation quality, with RIBOSPAN achieving the strongest overall performance across diverse RNA types and retaining a clear advantage on long RNAs. Building on the same backbone, we develop a multidimensionally conditioned discrete-diffusion framework for full-length mRNA generation and redesign, including synonymous-codon diffusion for protein-preserving CDS optimization. Together, RIBOSPAN establishes a powerful long-context foundation for transferable RNA representation learning and full-transcript mRNA design.

Interpretability research increasingly asks when concepts emerge during training and whether linear probes recover real structure, but in language models these claims are hard to validate because language offers little ground-truth ordering of concepts or relationships among them. We propose the use of astronomical ground truth through AstroPT, a transformer trained on millions of galaxy images, as a calibration testbed. AstroPT is an LLM-like model trained within a domain where the difficulty ordering of concepts and the relations among them are known in advance. Probing frozen representations across checkpoints, layers, model sizes, and objective choices, we find that galaxy properties emerge in a fixed order that tracks their known difficulty---quantities written almost directly into the pixels (band magnitude) become decodable early in training and shallow in the network, while multiband/spectra based and inferred quantities (such as redshift and specific star formation rate) emerge later and deeper. This order is invariant to our tested training objectives, and scales in magnitude but not in sequence with capacity. Our linear probe directions further recover the known physical structure among galaxy properties. Our findings suggest that astronomy offers a controlled sandbox for calibrating mechanistic interpretability methods we otherwise apply to LLMs blind.

Agent benchmarks often evaluate only final answers even when agents run on stateful runtimes. We argue this under-specifies what is being evaluated: the proper unit is a declared model-plus-runtime configuration whose failures can occur in evidence acquisition, runtime routing, safety boundaries, or repeated execution. We present ClawProBench, a trace-aware benchmark for runtime-native agent evaluation instantiated on OpenClaw, a live agent runtime with workspace tools and native surfaces for browsing, memory, messaging, scheduling, skills, and subagents. ClawProBench defines two tracks: a 102-scenario full profile with live workspace and native-runtime routing tasks, and a frozen 68-scenario holdout with closed-world JSON output contracts for robust ranking. Trials are scored from execution traces via a safety-gated formula combining correctness, process quality, and efficiency, preserving failure evidence for audit. Our anonymous artifact includes benchmark definitions, scoring code, manifests and sanitized traces. We evaluate 68 configurations on the full profile and 37 on holdout. The top safety-gated average trace score is 0.7671. Native-runtime tasks underperform workspace-live tasks (0.5238 vs. 0.6415). On holdout, pass@k-any outperforms strict three-trial pass (0.6638 vs. 0.2890), while full-profile and holdout rankings show weak alignment (Spearman 0.1300). Rankings based purely on correctness differ substantially from process-aware, safety-gated and strict-pass views. Final-answer leaderboards may hide native-surface weaknesses, one-off successes and trace-local agent failure modes.

Electrocardiogram (ECG) recordings are sensitive biomedical data, limiting the ability of hospitals and wearable devices to share raw signals for centralized model training. Federated learning addresses this practical privacy constraint by enabling collaborative model training while keeping raw biosignal data at their respective sources. However, federated ECG classification remains challenging due to limited client-side samples, imbalanced arrhythmia labels, and non-independent and identically distributed (non-IID) data across clients. These constraints require classifiers that are both communication-efficient and robust to cross-client distribution shifts. In this work, we evaluate a hybrid quantum-inspired Kolmogorov-Arnold network (HQKAN) against a multilayer perceptron (MLP) for five-class arrhythmia classification on the MIT-BIH dataset and three-class classification on the INCART dataset under federated averaging (FedAvg). Across multiple client configurations, HQKAN improves most aggregate and minority-class metrics while using 37.35% fewer trainable parameters and reducing communication cost by 24.89% on MIT-BIH; on INCART, it achieves corresponding reductions of 44.81% and 36.41%. These results indicate that HQKAN offers a compact, communication-efficient and robust alternative to the MLP baseline for privacy-aware federated learning on biosignal data.

Face presentation attack detection (PAD) is traditionally formulated as a face-specific problem, although many of the visual artifacts introduced by print, replay, and recapture processes are not inherently tied to facial appearance. In this work, we investigate whether transferable PAD representations can be learned without using faces during downstream PAD training. To this end, we introduce TPO, a controlled face-free presentation attack dataset consisting of bona fide, print, and replay recordings of, almost randomly chosen, tomatoes, potatoes, and onions acquired under protocols that closely mirror conventional face PAD datasets. Using a foundation-model-based PAD architecture, we demonstrate that a detector trained on TPO achieves an average AUC of 92.70% across four standard cross-dataset face PAD benchmarks, outperforming training on synthetic faces and remaining competitive with models trained on real face datasets. Conversely, models trained on face PAD datasets transfer consistently above chance to TPO, suggesting that the learned representations capture characteristics of the presentation process rather than object semantics. Furthermore, incorporating TPO into conventional face PAD training consistently improves cross-dataset performance under fixed optimization budgets, indicating that face-free data provides complementary information rather than simply additional training samples. Finally, representation and frequency analyses provide further evidence that transferable PAD representations cannot be explained by a single spectral artifact but instead encode richer presentation cues shared across object categories. Together, these results provide empirical evidence that transferable presentation attack representations can be learned independently of facial content, opening new opportunities for privacy-preserving and identity-independent PAD development.

Robot policies receive heterogeneous observations at each decision step, yet sequence models differ in how they organize these inputs over time. We introduce WorldToken, a time-first policy instantiation that fuses multiview images, proprioception, and task conditioning within each policy timestep into one world token. A causal temporal Transformer models the resulting world-token sequence, and a diffusion action head generates action chunks. On 23 RoboCasa tasks, an 85.3M-parameter policy trained from scratch apart from a frozen pretrained CLIP text encoder achieves 59.45% mean closed-loop success using 2,900 generated demonstrations per task. A complete factorial sweep over five dataset sizes, five model sizes, and two training seeds shows consistent gains from additional target-domain data and diminishing returns beyond moderate model size. Under same-checkpoint history truncation, reducing visible history to one or two policy timesteps lowers closed-loop success for all 50 RoboCasa policies. On RMBench Blocks Ranking, reducing visible history from 146 to 8 seconds lowers evaluator success from 95% to 28%, while an exploratory extended rollout sustains the reference swap sequence for over 850 seconds. These results establish the empirical feasibility of the complete WorldToken instantiation and characterize its data-scaling and temporal-context behavior under the tested recipes. They do not establish superiority over alternative sequence organizations or isolate which components of the complete implementation drive the observed performance.

Deep face recognition (FR) models reach near-saturated accuracy but remain opaque: a practitioner cannot ask which semantic attributes a similarity score relied upon. EXPL-FR answers this inside the FR model's own embedding space. A lightweight adapter aligns a vision-language model's (VLM) image encoder with the frozen FR space, trained on face images alone and never on text. Because the VLM's encoders share one space, the same adapter applies to the text encoder, turning 978 attribute prompts in 22 categories, also extendable, into FR-space anchors at no extra cost. We do not assume this transfer works: a face-verification protocol measures it, and an ablation changing only the adapter isolates its contribution. Not every concept survives, because an FR model earns its invariances by discarding the factors it must verify identities across. A label-free detectability measure compares each concept's separability in FR space against the VLM space, and the 100 most detectable form the model's readable semantic signature, which separates identities better than the full vocabulary. We cover four FR backbones and two VLM encoders, EXPL-FR needs no architecture access, and supports identity-level, per-image, and differential explanations. We benchmark attribute-level auditing under three supervision settings, human labels (current practice), VLM pseudo-labels, and our fully prompt-driven audit, against real verification behavior. With no labels, the prompt-driven audit ranks four FR models by their measured per-ethnicity RFW errors and ranks controlled attribute changes by their true verification cost.