7 papers
Off-Context GRPO: Learning to Reason on Hard Problems using Privileged Information
Priyank Agrawal, Ankur Samanta, Shervin Ghasemlou +4
Reinforcement learning with verifiable rewards (RLVR) improves reasoning in large language models. Yet, typical RLVR approaches fail on difficult problems: when a model cannot gene…
BayesBench: Evaluating LLM Belief Trajectories Under Multi-Turn Evidence Accumulation
Ankur Samanta, Akshayaa Magesh, Tal Lancewicki +7
Large language models (LLMs) are typically deployed in multi-turn conversations, where each turn provides new evidence that should reduce epistemic uncertainty about their environm…
Structure Enables Effective Self-Localization of Errors in LLMs
Ankur Samanta, Akshayaa Magesh, Ayush Jain +8
Self-correction in language models remains elusive. In this work, we explore whether language models can explicitly localize errors in incorrect reasoning, as a path toward buildin…
Credit Assignment with Resets in Language Model Reasoning
Ankur Samanta, Akshayaa Magesh, Ayush Jain +7
Contemporary reinforcement learning with verifiable reward methods post-train language models on multi-step reasoning by assigning a single outcome reward uniformly across all toke…
Self-Improvement of Language Models by Post-Training on Multi-Agent Debate
Ankur Samanta, Akshayaa Magesh, Runzhe Wu +7
Self-improvement, where models improve beyond their current performance without external supervision, remains a challenge. The core difficulty is sourcing a training signal stronge…
Improving Generative Ad Text on Facebook using Reinforcement Learning
Daniel R. Jiang, Alex Nikulkov, Yu-Chia Chen +2
Generative artificial intelligence (AI), in particular large language models (LLMs), is poised to drive transformative economic change. LLMs are pre-trained on vast text data to le…