finding
active
finding:sae-emotion-subspace-overlap-correlates-with-variance-residualized-persistence-in-cogito-spearman-0-413-p-4-4e-196SAE emotion subspace overlap correlates with variance-residualized persistence in Cogito: Spearman +0.413, p = 4.4e-196.
Strong positive relationship between emotion alignment and SAE feature persistence in Cogito
Source paper
extracted_fromScott Sauers · Imago · Janus · Antra Tessera
Neighborhood — ranked by edge-count
Claims (1)
claim
- Novel finding that agentic self-evaluation of emotionality correlates with feature persistence
Findings (1)
finding
- Demonstrates that SAE features more aligned with the emotion subspace are more persistent in Cogito after variance control
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.
- Emotion probe persistence correlation of 0.214 in Cogito v2.1 vs 0.099 for random vectorsfinding0.830Quantifies emotion feature persistence above random baseline in Cogito across 240 multi-turn conversations
- Explains why variance correction is needed to see the self-evaluation–persistence relationship
- Demonstrates emotion-specific persistence beyond variance effects in Cogito
- Demonstrates that Cogito emotion probes are persistently active beyond what is explained by their variance alone
- Quantitative measure of emotion feature persistence vs random baseline in Cogito
- Shows self-evaluated emotionality is negatively confounded by variance, requiring variance control to reveal the true signal
- Shows that model self-report of emotion predicts long-range feature persistence
- Fraction of an SAE feature's length lying inside the 171-dimensional subspace spanned by emotion probes, computed via SVD orthogonalization
Restated by (1)
cosine ≥ 0.90Other entities that say roughly the same thing. May be merge candidates or independent restatements across papers.