Juntao Li

Revisiting Long-context Modeling from Context Denoising Perspective

$Re^{2}$: Unlocking LLM Reasoning via Reinforcement Learning with Re-solving

LongRM: Revealing and Unlocking the Context Boundary of Reward Modeling

FAPO: Flawed-Aware Policy Optimization for Efficient and Reliable Reasoning

CaliDrop: KV Cache Compression with Query-based Calibration

Where Matters More Than What: Decoding-aligned KV Cache Compression via Position-aware Pseudo-queries

Training-Free Self-Scheduling for Efficient LLM Inference Serving

Self-Healing: Recovering Pruned Large Reasoning Models via Reinforcement Learning

LongFlow: Efficient KV Cache Compression for Reasoning Models

BatonVoice: An Operationalist Framework for Enhancing Controllable Speech Synthesis with Linguistic Intelligence from LLMs

GRM-Omni: Generative Omni-modality Reward Modeling via Meta Reward Learning

SCAN: Self-Denoising Monte Carlo Annotation for Robust Process Reward Learning

Thoughts Are All Over the Place: On the Underthinking of Long Reasoning Models

Revealing and Mitigating Over-Attention in Knowledge Editing

Scaling Code-Assisted Chain-of-Thoughts and Instructions for Model Reasoning

Beware of Calibration Data for Pruning Large Language Models

LOGO --- Long cOntext aliGnment via efficient preference Optimization

Unleashing Reasoning Capability of LLMs via Scalable Question Synthesis from Scratch

LOGO --- Long cOntext aliGnment via efficient preference Optimization

InfCycle: Learning to Use Tools via Inference Compute and Cycle Consistency

SinkQ: Accurate 2-bit KV Cache Quantization with Dynamic Sink Tracking

Improving Rationality in the Reasoning Process of Language Models through Self-playing Game

Thoughts Are All Over the Place: On the Underthinking of o1-Like LLMs

Path Selection Makes BERT-family Good Generators

L-CiteEval: Do Long-Context Models Truly Leverage Context for Responding?

CASD: Enhancing Generation Accuracy via Context-Aware Speculative Decoding