finding
active
finding:steering-effect-size-across-35-emotion-vectors-correlates-with-probe-preference-correlation-at-r-0-85Steering effect size across 35 emotion vectors correlates with probe-preference correlation at r=0.85
Emotions correlating with preference causally drive it.
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
- Key finding.
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.
- Steering effect between concepts correlates with My centroid distance, r=.65 (p<.001) for Emotionsfinding0.831Geometry predicts steering.
- Shows that causal steering effects persist over long ranges for a substantial fraction of emotion probes
- Emotion probe persistence correlation of 0.214 in Cogito v2.1 vs 0.099 for random vectorsfinding0.814Quantifies emotion feature persistence above random baseline in Cogito across 240 multi-turn conversations
- Evidence of a bottleneck between richer internal variation and final report distribution in impulsivity→interest condition
- Weaker geometry-steering correlation.
- Finding that the two evaluation modalities frequently diverge in their interpretation of the same SAE feature
- Validates that steering vectors capture reflection semantics by finding tokens reported in related work.
- Linear probes constructed to measure 171 emotion concepts in model activations with surface semantic content removed