Sotabase

Career

· Research Scientist, Meta2014–
· Research Scientist, Google2009–2014
· Research Staff Member, NEC Labs America2003–2009
· Researcher, Biowulf Technologies2000–2001
· Research Scientist, Max Planck Institute for Biological Cybernetics2000–2003
· Visiting Research Professor, New York University (NYU)
· Ph.D. in Machine Learning, Royal Holloway, University of London

Publications (95)

Self-Rewarding Language Models
International Conference on Machine Learning · 2024
489
cited
Chain-of-Verification Reduces Hallucination in Large Language Models
Annual Meeting of the Association for Computational Linguistics · 2023
345
cited
Training Large Language Models to Reason in a Continuous Latent Space
arXiv.org · 2024
335
cited
Iterative Reasoning Preference Optimization
Neural Information Processing Systems · 2024
207
cited
Meta-Rewarding Language Models: Self-Improving Alignment with LLM-as-a-Meta-Judge
Conference on Empirical Methods in Natural Language Processing · 2024
162
cited
Distilling System 2 into System 1
arXiv.org · 2024
131
cited
Some things are more CRINGE than others: Preference Optimization with the Pairwise Cringe Loss
arXiv.org · 2023
109
cited
Branch-Train-MiX: Mixing Expert LLMs into a Mixture-of-Experts LLM
arXiv.org · 2024
90
cited
System 2 Attention (is something you might need too)
arXiv.org · 2023
82
cited
Self-Taught Evaluators
arXiv.org · 2024
55
cited
Contextual Position Encoding: Learning to Count What's Important
arXiv.org · 2024
53
cited
SWEET-RL: Training Multi-Turn LLM Agents on Collaborative Reasoning Tasks
arXiv.org · 2025
53
cited
Reverse Training to Nurse the Reversal Curse
arXiv.org · 2024
51
cited
NaturalReasoning: Reasoning in the Wild with 2.8M Challenging Questions
arXiv.org · 2025
47
cited
Thinking LLMs: General Instruction Following with Thought Generation
arXiv.org · 2024
43
cited
Following Length Constraints in Instructions
Conference on Empirical Methods in Natural Language Processing · 2024
35
cited
Diverse Preference Optimization
arXiv.org · 2025
29
cited
Proceedings of the 1st Workshop on Representation Learning for NLP
2016
26
cited
TOOLVERIFIER: Generalization to New Tools via Self-Verification
Conference on Empirical Methods in Natural Language Processing · 2024
24
cited
Self-Consistency Preference Optimization
International Conference on Machine Learning · 2024
23
cited
Sotabase
Jason Weston | Researcher Profile | Sotabase | Sotabase