HuggingFace
Self-Evolving Coding Agents
Large language models are increasingly embedded in software engineering workflows as coding agents that can inspect r...
🤖 AI 前沿资讯
每天自动聚合 AI 领域最新动态
HuggingFace
Are the Financial Reasoning from LLMs Credible? A Real World Test over Long-Horizon Statements
Do Large Language Models (LLMs) possess genuine structural reasoning, or merely rely on surface-level pattern matchin...
HuggingFace
Multi-Task Multi-Frame Visual Piano Transcription
Audio-based piano transcription performs well on onset, pitch, and velocity, but the sustain pedal lets sound persist...
HuggingFace
When Attention Goes Blind: Numerical Failure in ALiBi Positional Encodings
We identify a previously overlooked failure mode of ALiBi positional encoding: its linear bias scaling underflows flo...
HuggingFace
Better, Stronger, Faster, and Broader: Structured All-Mask Prediction for MLLM-Based Segmentation
MLLM-based segmentation faces a core segmentation trilemma: high segmentation performance, preserved dialogue ability...
HuggingFace
ChronoLens: Measuring Language Change Across Time, Languages, and Linguistic Levels
Historical language change affects morphology, syntax, semantics, and pragmatics, yet computational studies typically...
HuggingFace
RestoreKV: Recovering Full-Cache Behavior Under Aggressive Query-Agnostic KV Cache Eviction
Query-agnostic KV cache eviction compresses a context once and reuses the resulting cache for arbitrary future querie...
HuggingFace
CURV: Enhancing Chart Understanding Through Curriculum Visual Grounded Reasoning
Chart question answering (CQA) requires multimodal large language models (MLLMs) to integrate visual comprehension wi...
HuggingFace
ReflectRL: Learning from Golden Negative Trajectories via Reflective-to-Direct Reasoning
On-policy training has emerged as a powerful post-training paradigm for improving the reasoning capabilities of large...
HuggingFace
LegalPincite: Multi-level Legal Information Retrieval Dataset
A common task in legal Information Retrieval (IR) is to find relevant legal sources from case-law collections. While ...
HuggingFace
ARCHead: Activation-Metric Residual Correction for Large Language Model Output Heads
Weight-only quantization substantially reduces the storage of large language model (LLM) transformer blocks, but prac...
HuggingFace
When Many Answers Are Valid, Voting Fails: Symbolic Verification for Best-of-K Causal Reasoning in LLMs
Self-consistency assumes the most frequent answer among sampled reasoning traces is the most reliable, but this can f...