AI研究論文每日精選

Large Language Models and Mathematical Reasoning Failures

EQ-VAE: Equivariance Regularized Latent Space for Improved Generative Image Modeling

Dyve: Thinking Fast and Slow for Dynamic Process Verification

Better Embeddings with Coupled Adam

The Mirage of Model Editing: Revisiting Evaluation in the Wild

Towards Data-Efficient Pretraining for Atomic Property Prediction

ReLearn: Unlearning via Learning for Large Language Models

Diffusion Models without Classifier-free Guidance

Memory, Benchmark & Robots: A Benchmark for Solving Complex Tasks with Reinforcement Learning

Language Complexity Measurement as a Noisy Zero-Shot Proxy for Evaluating LLM Performance

Show Me the Work: Fact-Checkers' Requirements for Explainable Automated Fact-Checking

Diffusion-Sharpening: Fine-tuning Diffusion Models with Denoising Trajectory Sharpening

SAFE-SQL: Self-Augmented In-Context Learning with Fine-grained Example Selection for Text-to-SQL

Explorer: Scaling Exploration-driven Web Trajectory Synthesis for Multimodal Web Agents

Building A Proof-Oriented Programmer That Is 64% Better Than GPT-4o Under Data Scarsity

One Example Shown, Many Concepts Known! Counterexample-Driven Conceptual Reasoning in Mathematical LLMs

PhysReason: A Comprehensive Benchmark towards Physics-Based Reasoning

video-SALMONN-o1: Reasoning-enhanced Audio-visual Large Language Model

Cuckoo: An IE Free Rider Hatched by Massive Nutrition in LLM's Nest

Ask in Any Modality: A Comprehensive Survey on Multimodal Retrieval-Augmented Generation

I Think, Therefore I Diffuse: Enabling Multimodal In-Context Reasoning in Diffusion Models

How Do LLMs Acquire New Knowledge? A Knowledge Circuits Perspective on Continual Pre-Training

System Message Generation for User Preferences using Open-Source Models

Learning Getting-Up Policies for Real-World Humanoid Robots

ILIAS: Instance-Level Image retrieval At Scale

Native Sparse Attention: Hardware-Aligned and Natively Trainable Sparse Attention

CRANE: Reasoning with constrained LLM generation

IHEval: Evaluating Language Models on Following the Instruction Hierarchy

Sailor2: Sailing in South-East Asia with Inclusive Multilingual LLMs

Can a Single Model Master Both Multi-turn Conversations and Tool Use? CALM: A Unified Conversational Agentic Language Model

ExaGPT: Example-Based Machine-Generated Text Detection for Human Interpretability

Data Valuation using Neural Networks for Efficient Instruction Fine-Tuning

Intuitive physics understanding emerges from self-supervised pretraining on natural videos

Talk Structurally, Act Hierarchically: A Collaborative Framework for LLM Multi-Agent Systems

SWE-Lancer: Can Frontier LLMs Earn $1 Million from Real-World Freelance Software Engineering?

SURGE: On the Potential of Large Language Models as General-Purpose Surrogate Code Executors

MagicArticulate: Make Your 3D Models Articulation-Ready

HermesFlow: Seamlessly Closing the Gap in Multimodal Understanding and Generation