hypothesis
active
hypothesis:when-mechanisms-agree-they-overwhelmingly-drive-model-behaviorWhen mechanisms agree they overwhelmingly drive model behavior
Tested in D.1.
Source paper
extracted_from(2025) · Yoav Gur-Arieh · Mor Geva · Atticus Geiger
Neighborhood — ranked by edge-count
Papers (1)
paper
Findings (1)
finding
- Mechanism agreement.
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.
- Context length ablation.
- Central claim motivating the product.
- Speculation about false positives.
- Central interpretive claim organizing the entire paper's results
- The method can steer the model in both positive and negative directions on the target semantic.
- Directed modulation.
- Practical implication.
- How general are the model's introspective mechanisms? Do they have a global representation of thoughts?question0.776Question about uniformity of introspection mechanisms.