quote
active
quote:steering-operates-by-changing-concept-priors-while-in-context-learning-leads-to-an-accumulation-of-evidencesteering operates by changing concept priors, while in-context learning leads to an accumulation of evidence
Core mechanism statement.
Source paper
extracted_from(2025) · Eric Bigelow · Daniel Wurgaft · YingQiao Wang · Noah D. Goodman +3
Neighborhood — ranked by edge-count
Papers (1)
paper
Claims (1)
claim
- Mechanistic split.
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.
- Addresses skeptical alternative that reports reflect only conversational content
- Nuanced interpretive claim about the limits of steering as a mechanism for reflection enhancement.
- Central interpretive claim organizing the entire paper's results
- Extension of the thesis to deployed LLM inference via in-context learning
- Mechanism claim supported by transcript analysis and the fact that the steering vector was extracted from a model that never writes type hints.
- Demonstration of steering reasoning length.
- Explanation of brittle steering.
- Suggests localization.