institute
active
institute:anthropic-alignment-science-teamAnthropic Alignment Science Team
Authors' affiliation.
Neighborhood — ranked by edge-count
Thinkers (4)
thinker
- Ethan Perezaffiliated_with
- Jared Kaplanaffiliated_with
- Jan Leikeaffiliated_with
- Yanda Chenaffiliated_with
Papers (1)
paper
Concepts (1)
concept
- Claude 3.7 Sonnetaffiliated_withAnthropic model tested in Experiments 1, 3, 4; shows 100% experience reporting under self-referential induction
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- The team that produced this developing-work update.
- Field within which this work has implications for evaluating alignment progress.
- The goal of making model behavior match human values and intentions, often addressed during post-training.
- Lab behind Claude models and Constitutional AI training approach; represents highest baseline scores and lowest prompt lift.
- A learnable invertible transformation in DAS that maps neural representations to a basis aligned with causal variables
- Measure of similarity between the similarity structures (kernels) induced by two different representations
- AI alignment organization developing interpretability methods relevant to consciousness detection