Beidi Chen

FSA: An Alternative Efficient Implementation of Native Sparse Attention Kernel

STEM: SCALING TRANSFORMERS WITH EMBEDDING MODULES

Prosperity before Collapse: How Far Can Off-Policy RL Reach with Stale Data on LLMs?

WWW.Serve: A Decentralized Framework for Collaborative LLM Serving

Jackpot: Align Actor-Policy Distribution for scalable and stable RL for LLM

Scalable LLM Math Reasoning Acceleration with Low-rank Distillation

CamoDocs: Poisoning Attack against Retrieval-Augmented Language Models

Kinetics: Rethinking Test-Time Scaling Law

Multiverse: Your Language Models Secretly Decide How to Parallelize and Merge Generation

MagicPIG: LSH Sampling for Efficient LLM Generation

ShadowKV: KV Cache in Shadows for High-Throughput Long-Context LLM Inference

MagicDec: Breaking the Latency-Throughput Tradeoff for Long Context Generation with Speculative Decoding

ShadowKV: KV Cache in Shadows for High-Throughput Long-Context LLM Inference

APE: Faster and Longer Context-Augmented Generation via Adaptive Parallel Encoding

GSM-$\infty$: How Do your LLMs Behave over Infinitely Increasing Reasoning Complexity and Context Length?

Act Only When It Pays: Efficient Reinforcement Learning for LLM Reasoning via Selective Rollouts

Memory Mosaics

Zeroth-Order Fine-Tuning of LLMs with Transferable Static Sparsity

FACTOR: Factoring Complexity and Context Length in Long-Context Model Evaluation

Speculative Prefill: Turbocharging TTFT with Lightweight and Training-Free Token Importance Estimation

It Takes Two: On the Seamlessness between Reward and Policy Model in RLHF