finding
active
finding:post-training-shifts-correlate-across-neutral-and-challenging-scenarios-r-0-90Post-training shifts correlate across neutral and challenging scenarios (r=0.90)
Context-independent transformation.
Source paper
extracted_from(2026) · Nicholas Sofroniew · Isaac Kauvar · William Saunders · Runjin Chen +12
Neighborhood — ranked by edge-count
Papers (1)
paper
Claims (1)
claim
- Interpretation of shifts.
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.
- Central interpretive claim and motivation for future work
- How does different post-training data shift a model's position along persona dimensions?question0.794Future work direction: using persona space to study effects of training data on model character
- Demonstrates that persona vectors capture trait-specific signal beyond general misalignment signal
- Finding that base models have high false positives and no net positive performance.
- Broader research area: methods to align model behavior after initial training, where undesired behaviors can emerge.
- Assertion about the role of post-training in eliciting introspection.
- Load-bearing summary of the paper's core finding about persona stability
- Preference circuitry established in pretraining.