14 papers
How Reinforcement Learning After Next-Token Prediction Facilitates Learning
Nikolaos Tsilivis, Eran Malach, Karen Ullrich +1
Recent advances in reasoning domains with neural networks have primarily been enabled by a training recipe that optimizes Large Language Models, previously trained to predict the n…
Asymmetric REINFORCE for off-Policy Reinforcement Learning: Balancing positive and negative rewards
Charles Arnal, Gaëtan Narozniak, Vivien Cabannes +3
Reinforcement learning (RL) is increasingly used to align large language models (LLMs). Off-policy methods offer greater implementation simplicity and data efficiency than on-polic…
OpenApps: Simulating Environment Variations to Measure UI-Agent Reliability
Karen Ullrich, Jingtong Su, Claudia Shi +7
Reliability is key to realizing the promise of autonomous UI-Agents, multimodal agents that directly interact with apps in the same manner as humans, as users must be able to trust…
Don't Waste Mistakes: Leveraging Negative RL-Groups via Confidence Reweighting
Yunzhen Feng, Parag Jain, Anthony Hartshorn +2
Reinforcement learning with verifiable rewards (RLVR) has become a standard recipe for improving large language models (LLMs) on reasoning tasks, with Group Relative Policy Optimiz…
What Characterizes Effective Reasoning? Revisiting Length, Review, and Structure of CoT
Yunzhen Feng, Julia Kempe, Cheng Zhang +2
Large reasoning models (LRMs) spend substantial test-time compute on long chain-of-thought (CoT) traces, but what *characterizes* an effective CoT remains unclear. While prior work…
Flavors of Margin: Implicit Bias of Steepest Descent in Homogeneous Neural Networks
Nikolaos Tsilivis, Eitan Gronich, Julia Kempe +1
We study the implicit bias of the general family of steepest descent algorithms with infinitesimal learning rate in deep homogeneous neural networks. We show that: (a) an algorithm…