Contributor Publications
Character Training for Risk-Averse Agents
Arav Dhoot*, Punya Syon Pandey*, Jamie Johnson, Daniel Tan, Elliott Thornley, David Africa, “Character Training for Risk-Averse Agents,” International Conference on Learning Representations (ICLR) 2027 (under review)
Fictionalism about Personas: Folk Psychology as an Interpretability Strategy
Weiming Sheng, “Fictionalism about Personas: Folk Psychology as an Interpretability Strategy,” Proceedings of the 43 rd International Conference on Machine Learning, Seoul, South Korea, PMLR 306, 2026
Subtly Wrong Advice Misaligns Models More Than Obviously Wrong Advice
Arav Dhoot, C. Corry, “Subtly Wrong Advice Misaligns Models More Than Obviously Wrong Advice,” LESSWRONG 2026
Predicting how Model Propensities Generalize out of Distribution
June Hunter, Lily Wen, Owen Terry, Johannes Taraz, Ben Cohen, Niels Warncke, “Predicting how model propensities generalize out of distribution,” arXiv:2610.14937
Measuring Alignment Drift via Trajectory Prefixes
Owen Terry, Maksym Andriushchenko, “Measuring alignment drift via trajectory prefixes,” LESSWRONG 2026
LLM CoTs Remain Monitorable When Being Unfaithful Requires Computation: Replicating and Extending the Necessity Effect Across Eleven Models
Arav Dhoot, Yixiong Hao, Zephaniah Roe, “LLM CoTs Remain Monitorable When Being Unfaithful Requires Computation: Replicating and Extending the Necessity Effect Across Eleven Models,” LESSWRONG 2026
Relational Structural Causal Models
Adiba Ejaz, Elias Bareinboim, “Relational Structural Causal Models,” Proceedings of the 43rd International Conference on Machine Learning (ICML) 2026
Implications of Continual Learning for LLM Agents: Introduction
RohanS, Rauno Arike, Owen Terry, Achu Menon, Zhijing Jin, Francis Rhys Ward, Seth Herd, “Implications of Continual Learning for LLM Agents: Introduction,” LESSWRONG 2026
Failing to Ragebait the New Gemma
Arav Dhoot, Neil Shah, David Africa, “Failing to Ragebait the New Gemma,” LESSWRONG 2026
Consistency Training Along the Transformer Stack
Sukrati Gautam*, Neil Shah*, Arav Dhoot*, Bryan Maruyama*, Caroline Wei* (and others), “Consistency Training Along the Transformer Stack,” AI4GOOD Workshop @ International Conference on Machine Learning (ICML) 2026 & EMNLP 2026
Attention-Based Neural Causal Models Under Markov Equivalence
Adiba Ejaz*, Yushu Pan*, Hongshuo Yang*, Elias Bareinboim, “Attention-Based Neural Causal Models Under Markov Equivalence,” Proceedings of the 40th Annual Conference on Neural Information Processing Systems (NeurIPS) 2026
Building a Correct-by-Design Lakehouse. Data Contracts, Versioning, and Transactional Pipelines for Humans and Agents
Weiming Sheng, Jinlang Wang, Manuel Barros, Aldrin Montana, Jacopo Tagliabue, Luca Bigon, “Building a Correct-by-Design Lakehouse. Data Contracts, Versioning, and Transactional Pipelines for Humans and Agents,” ICML 2026 Workshop (poster)