hypothesis
active
hypothesis:r1-implicitly-recognizes-confusion-or-incoherence-when-activations-are-overly-perturbed-causing-it-to-stop-and-course-correct

R1 implicitly recognizes confusion or incoherence when activations are overly perturbed, causing it to stop and course-correct

Working hypothesis for the oversteering reversion.

Source paper

extracted_from
Under the Hood of a Reasoning Model - Goodfire
Dron Hazra

Neighborhood — ranked by edge-count

Findings (1)

finding

Concepts (1)

concept

Hypotheses (1)

hypothesis

Related by similarity (8)

cosine ≥ 0.65 · no typed edge

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.