← Search

Ethan Hall

2 accepted papers

2024

RLAIF vs. RLHF: Scaling Reinforcement Learning from Human Feedback with AI Feedback

ICML 2024poster

Reinforcement learning from human feedback (RLHF) has proven effective in aligning large language models (LLMs) with human preferences, but gathering high-quality preference labels is expensive. RL from AI Feedback (RLAIF), introduced in Bai et al. (2022b), offers a promising alternative that trains…

Cited by 99SourcePDFScholar
2024

Understanding Transformer Reasoning Capabilities via Graph Algorithms

NeurIPS 2024poster

Which transformer scaling regimes are able to perfectly solve different classes of algorithmic problems? While tremendous empirical advances have been attained by transformer-based neural networks, a theoretical understanding of their algorithmic reasoning capabilities in realistic parameter regim…

Cited by 23SourcePDFScholar