11 papers
Simultaneous Multi-objective Alignment Across Verifiable and Non-verifiable Rewards
Yiran Shen, Yu Xia, Jonathan Chang +1
Aligning large language models to human preferences is inherently multidimensional, yet most pipelines collapse heterogeneous signals into a single objective. We seek to answer wha…
Introspective X Training: Feedback Conditioning Improves Scaling Across all LLM Training Stages
Brandon Cui, Ximing Lu, Jaehun Jung +7
We tackle the question of how to scale more efficiently across the many, ever-growing stages of current LLM training pipelines. Our guiding intuition stems from the fact that the d…
How to Instruct Your Robot: Dense Language Annotations Power Robot Policy Learning
Bosung Kim, Ruiyi Wang, David Acuna +5
Scaling robot policy learning is bottlenecked by the cost of collecting demonstrations, while language annotations for existing demonstrations are comparatively cheap. We study lan…
MASS-DPO: Multi-negative Active Sample Selection for Direct Policy Optimization
Rohan Surana, Xintong Li, Sheldon Yu +7
Multi-negative preference optimization under the Plackett--Luce (PL) model extends Direct Preference Optimization (DPO) by leveraging comparative signals across one preferred and m…
Long Grounded Thoughts: Synthesizing Visual Problems and Reasoning Chains at Scale
David Acuna, Chao-Han Huck Yang, Yuntian Deng +6
Despite rapid progress, multimodal reasoning still lacks a systematic approach to synthesize large-scale vision-centric datasets beyond visual math. We introduce a framework able t…
Golden Goose: A Simple Trick to Synthesize Unlimited RLVR Tasks from Unverifiable Internet Text
Ximing Lu, David Acuna, Jaehun Jung +12
Reinforcement Learning with Verifiable Rewards (RLVR) has become a cornerstone for unlocking complex reasoning in Large Language Models (LLMs). Yet, scaling up RL is bottlenecked b…