Tristan Hume
Source-listed: Lead · Anthropic
Record snapshot: Sep 3, 2026. Coverage and affiliations may be incomplete or historical.
Selected work
5- Constitutional AI: Harmlessness from AI FeedbackSaved credit: Contributor · 2022Explore people connected to this work →
- Training a Helpful and Harmless Assistant with Reinforcement Learning from Human FeedbackSaved credit: Contributor · 2022Explore people connected to this work →
- Measuring Faithfulness in Chain-of-Thought ReasoningSaved credit: Contributor · 2023Explore people connected to this work →
- Scaling Monosemanticity: Extracting Interpretable Features from Claude 3 SonnetSaved credit: Contributor · 2026Explore people connected to this work →
- magic-traceSaved credit: Contributor · 2022Explore people connected to this work →