Brian R. Bartoldson

Researcher, Lawrence Livermore National Labs

4 papers at NeurIPS 2025

Homepage· OpenReview· Semantic Scholar· Google Scholar

Poster Session 1

1 paper

Wednesday, December 3, 2025 · 11:00 AM → 2:00 PM

Exhibit Hall C,D,E

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

#305 · Haizhong Zheng, Yang Zhou, Brian R. Bartoldson, Bhavya Kailkhura, Fan Lai, Jiawei Zhao, Beidi Chen

Poster Session 5

2 papers

Friday, December 5, 2025 · 11:00 AM → 2:00 PM

Exhibit Hall C,D,E

Scaling up Test-Time Compute with Latent Reasoning: A Recurrent Depth Approach

#3615 Spotlight · Jonas Geiping, Sean Michael McLeish, Neel Jain, John Kirchenbauer, Siddharth Singh, Brian R. Bartoldson, Bhavya Kailkhura, Abhinav Bhatele, Tom Goldstein

We show that recurrent-depth transformers can be scaled to be effective language models, with particularly strong gains through additional compute for reasoning tasks.

The Common Pile v0.1: An 8TB Dataset of Public Domain and Openly Licensed Text

#102 · Nikhil Kandpal, Brian Lester, Colin Raffel, Sebastian Majstorovic, Stella Biderman, Baber Abbasi, Luca Soldaini, Enrico Shippole, A. Feder Cooper, Aviya Skowron, Shayne Longpre, Lintang Sutawika, Alon Albalak, Zhenlin Xu, Guilherme Penedo, Loubna Ben allal, Elie Bakouch, John David Pressman, Honglu Fan, Dashiell Stander, Guangyu Song, Aaron Gokaslan, John Kirchenbauer, Tom Goldstein, Brian R. Bartoldson, Bhavya Kailkhura, Tyler Murray

We collect an 8 TB of public domain and openly licensed text and use it to pre-train a performant 7B-parameter LLM.

Poster Session 6

1 paper

Friday, December 5, 2025 · 4:30 PM → 7:30 PM

Exhibit Hall C,D,E

Trajectory Balance with Asynchrony: Decoupling Exploration and Learning for Fast, Scalable LLM Post-Training

#304 · Brian R. Bartoldson, Siddarth Venkatraman, James Diffenderfer, Moksh Jain, Tal Ben-Nun, Seanie Lee, Minsu Kim, Johan Obando-Ceron, Yoshua Bengio, Bhavya Kailkhura

We improve the speed and performance of LLM post-training via a new asynchronous RL approach, leveraging an off-policy objective, replay buffer, and sampling strategies.