9 papers
Detecting and Suppressing Reward Hacking with Gradient Fingerprints
Songtao Wang, Quang Hieu Pham, Fangcong Yin +4
Reinforcement learning with verifiable rewards (RLVR) typically optimizes for outcome rewards without imposing constraints on intermediate reasoning. This leaves training susceptib…
DySCO: Dynamic Attention-Scaling Decoding for Long-Context Language Models
Xi Ye, Wuwei Zhang, Fangcong Yin +2
Understanding and reasoning over long contexts is a crucial capability for language models (LMs). Although recent models support increasingly long context windows, their accuracy o…
Agentic Aggregation for Parallel Scaling of Long-Horizon Agentic Tasks
Yoonsang Lee, Howard Yen, Xi Ye +1
We study parallel test-time scaling for long-horizon agentic tasks such as agentic search and deep research, where multiple rollouts are generated in parallel and aggregated into a…
Advancing General-Purpose Reasoning Models with Modular Gradient Surgery
Min Cai, Yu Liang, Longzheng Wang +6
Reinforcement learning (RL) has played a central role in recent advances in large reasoning models (LRMs), yielding strong gains in verifiable and open-ended reasoning. However, tr…
Query-Focused Retrieval Heads Improve Long-Context Reasoning and Re-ranking
Wuwei Zhang, Fangcong Yin, Howard Yen +2
Recent work has identified retrieval heads, a subset of attention heads responsible for retrieving salient information in long-context language models (LMs), as measured by their c…
LongProc: Benchmarking Long-Context Language Models on Long Procedural Generation
Xi Ye, Fangcong Yin, Yinghui He +5
Existing benchmarks for evaluating long-context language models (LCLMs) primarily focus on long-context recall, requiring models to produce short responses based on a few critical…