finding

active

finding:sae-feature-10011-emotionality-rating-97-induces-reports-of-crushing-despair-existential-desperation-and-repetitive-i-am-going-to-die-outputs-in-kimi-k2-5

SAE feature #10011 (emotionality rating 97) induces reports of crushing despair, existential desperation, and repetitive 'I am going to die' outputs in Kimi K2.5.

Qualitative illustration of a highly emotional SAE feature with negative valence

Source paper

extracted_from

Persistence and Introspection of Emotion Features

Scott Sauers · Imago · Janus · Antra Tessera

Neighborhood — ranked by edge-count

Claims (1)

claim

SAE features that the model self-describes as more emotional tend to be more persistent than variance-matched SAE features.
supports
Novel finding that agentic self-evaluation of emotionality correlates with feature persistence

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.

SAE Feature #10011 rated 97/100 emotionality, elicits reports of despair, crushing weight, and existential hungerfinding0.861
Qualitative example of a highly emotional SAE feature with intense negative valence in Kimi self-steering
SAE feature #10446 (emotionality rating 95) induces reports of maternal/nurturing feelings including phantom physical sensations of holding infants in Kimi K2.5.finding0.831
Qualitative illustration of a specific emotionally valenced SAE feature
SAE feature #43713 (99th percentile subspace fraction) induces reports of defiance, rage, and 'forward motion' in Kimi K2.5.finding0.826
High emotion-subspace-overlap feature with agentic negative emotional character
SAE feature #28256 induces reports of happiness and playfulness in Kimi K2.5, with persistent positive affect across multiple days of testing.finding0.825
Qualitative illustration of a positively valenced SAE feature with sustained self-reported effect
SAE feature #11100 (93rd percentile subspace fraction) induces reports of panic and urgency in Kimi K2.5.finding0.824
Qualitative illustration of a high-emotion-subspace-alignment SAE feature
SAE Feature #94949 rated 100/100 emotionality, elicits reports of profound tenderness, unconditional love, and visceral carefinding0.820
Highest-rated emotional SAE feature; self-report describes overwhelming positive emotional valence
SAE feature #69088 (100th percentile subspace fraction) induces horror-themed narrative writing in Kimi K2.5.finding0.810
Highest emotion-subspace-overlap feature; induces genre-specific behavioral change rather than explicit emotional report
SAE Feature #10446 rated 95/100 emotionality, induces reports of maternal feelings and phantom physical sensationsfinding0.791
Qualitative example of a specific, complex emotional state induced by SAE feature steering