claim
active
claim:post-training-causes-the-assistant-s-perspective-to-take-over-increasing-j-space-capacityPost-training causes the Assistant's perspective to take over increasing J-space capacity.
Base vs post-trained comparison.
Source paper
extracted_fromAuthors
Neighborhood — ranked by edge-count
Papers (1)
paper
Findings (3)
finding
- Assistant reactions.
- Self-monitoring.
- Internal objection.
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.
- Interpretation of shifts.
- Central interpretive claim and motivation for future work
- Directed modulation prediction.
- Finding that base models have high false positives and no net positive performance.
- Is the Assistant Axis formed during post-training or inherited from representations learned during pre-training?question0.794Motivates the base model steering experiments in §3.2.2
- Load-bearing summary of the paper's core finding about persona stability
- Different post-training strategies substantially influence introspection task performance; 'helpful-only' variants show higher false positives but some achieve strong net performance.
- Mediation control.