Thomas Icard
Source-listed: Full-time work on the science of interpretability (formal title not stated) · Goodfire
Record snapshot: Sep 3, 2026. Coverage and affiliations may be incomplete or historical.
Selected work
5- Holistic Evaluation of Language ModelsSaved credit: Contributor · 2022Explore people connected to this work →
- HELM (Holistic Evaluation of Language Models)Saved credit: MaintainerExplore people connected to this work →
- Causal Abstraction: A Theoretical Foundation for Mechanistic InterpretabilitySaved credit: Contributor · 2025Explore people connected to this work →
- Causal Distillation for Language ModelsSaved credit: Contributor · 2022Explore people connected to this work →
- Structuring Sparsity: Block-Sparse Featurizers Capture Visual Concept ManifoldsSaved credit: Contributor · 2026Explore people connected to this work →