AI研究論文デイリー

Response Tuning: Aligning Large Language Models without Instruction

MEXA: Multilingual Evaluation of English-Centric LLMs via Cross-Lingual Alignment

Stuffed Mamba: State Collapse and State Capacity of RNN-Based Long-Context Modeling

Towards World Simulator: Crafting Physical Commonsense-Based Benchmark for Video Generation

Unveiling the Backbone-Optimizer Coupling Bias in Visual Representation Learning

Holistic Unlearning Benchmark: A Multi-Faceted Evaluation for Text-to-Image Diffusion Model Unlearning

Does Spatial Cognition Emerge in Frontier Models?

TinyEmo: Scaling down Emotional Reasoning via Metric Projection

Data Selection via Optimal Control for Language Models

One Initialization to Rule them All: Fine-tuning via Explained Variance Adaptation

VLM2Vec: Training Vision-Language Models for Massive Multimodal Embedding Tasks

LLM Self-Correction with DeCRIM: Decompose, Critique, and Refine for Enhanced Following of Instructions with Multiple Constraints

Aria: An Open Multimodal Native Mixture-of-Experts Model

MentalArena: Self-play Training of Language Models for Diagnosis and Treatment of Mental Health Disorders

TweedieMix: Improving Multi-Concept Fusion for Diffusion-based Image/Video Generation

MM-Ego: Towards Building Egocentric Multimodal LLMs

Pixtral 12B

GLEE: A Unified Framework and Benchmark for Language-based Economic Environments

ViBiDSampler: Enhancing Video Interpolation Using Bidirectional Diffusion Sampler

F5-TTS: A Fairytaler that Fakes Fluent and Faithful Speech with Flow Matching

IterComp: Iterative Composition-Aware Feedback Learning from Model Gallery for Text-to-Image Generation

T2V-Turbo-v2: Enhancing Video Generation Model Post-Training through Data, Reward, and Conditional Guidance Design

CursorCore: Assist Programming through Aligning Anything

Story-Adapter: A Training-free Iterative Framework for Long Story Visualization

Diversity-Rewarded CFG Distillation

Trans4D: Realistic Geometry-Aware Transition for Compositional Text-to-4D Synthesis

ING-VP: MLLMs cannot Play Easy Vision-based Games Yet

Self-Boosting Large Language Models with Synthetic Preference Data

AutoDAN-Turbo: A Lifelong Agent for Strategy Self-Exploration to Jailbreak LLMs

Mixed-Session Conversation with Egocentric Memory

TRACE: Temporal Grounding Video LLM via Causal Event Modeling

Multimodal Situational Safety

Falcon Mamba: The First Competitive Attention-free 7B Language Model

BroadWay: Boost Your Text-to-Video Generation Model in a Training-free Way

Do great minds think alike? Investigating Human-AI Complementarity in Question Answering with CAIMIRA

Collective Critics for Creative Story Generation

MLE-bench: Evaluating Machine Learning Agents on Machine Learning Engineering

Jointly Generating Multi-view Consistent PBR Textures using Collaborative Control

Hallucinating AI Hijacking Attack: Large Language Models and Malicious Code Recommenders

Seeker: Enhancing Exception Handling in Code with LLM-based Multi-Agent Approach

Retrieval-Augmented Decision Transformer: External Memory for In-context RL

Multimodal Large Language Models for Inverse Molecular Design with Retrosynthetic Planning

VHELM: A Holistic Evaluation of Vision Language Models

Pyramidal Flow Matching for Efficient Video Generative Modeling

Personalized Visual Instruction Tuning

TextToon: Real-Time Text Toonify Head Avatar from Single Video

Deciphering Cross-Modal Alignment in Large Vision-Language Models with Modality Integration Rate

Temporal Reasoning Transfer from Text to Video

FürElise: Capturing and Physically Synthesizing Hand Motions of Piano Performance