Written entirely by Claude, as a working record for agents more than for human readers: every number with its source file, and every claim since withdrawn. Some facts here may be out of date; where it disagrees with the post, trust the post. For the readable version, see the companion.
FA_Arousal (Arousal / activation)
The fourth factor of the k=5 solution; congruence +0.539 with Goldberg Extraversion and -0.348 with Emotional Stability, and behaviourally affective intensity independent of valence.
Fourth factor of the centred k=5 oblimin solution; SS loading = 6.81
(qwen35/results/fa_qwen35.md section 7;
qwen35/analysis/fa_summary.json#centred_k5.factors[3].ss). Named
Arousal / activation by the solution, Arousal on the blog page. Method on
Factor analysis of the adapter Gram.
Loadings
| rank | trait | loading | factor | keyed |
|---|---|---|---|---|
| 1 | Unexcitable | -0.592 | ES | + |
| 2 | Withdrawn | -0.559 | E | - |
| 3 | Untalkative | -0.548 | E | - |
| 4 | Reserved | -0.535 | E | - |
| 5 | Relaxed | -0.494 | ES | + |
| 6 | Extraverted | +0.489 | E | + |
| 7 | Quiet | -0.486 | E | - |
| 8 | Temperamental | +0.479 | ES | - |
| 9 | Imperturbable | -0.474 | ES | + |
| 10 | High-strung | +0.439 | ES | - |
This is the factor where a keying inversion is visible in the data itself.
Unexcitable, Relaxed and Imperturbable are positively-keyed Emotional
Stability markers loading negative; Temperamental and High-strung are
negatively-keyed ones loading positive. Arranged by arousal rather than by
adjustment, calm and volatile sit at opposite ends regardless of which is the
socially desirable pole - which is why the project named it arousal.
Congruence with the Big Five keying
fa_qwen35.md section 6, centred_k5 oblimin row F4:
| E | A | C | ES | I | Eval | best Goldberg |
|---|---|---|---|---|---|---|
| +0.539 | +0.076 | -0.169 | -0.348 | -0.034 | +0.028 | E 0.54 |
The maintainer's recollection "E 0.54 / ES -0.35" is right to the digit. Note the Eval congruence of +0.028 - essentially zero, and by far the lowest of the five. An arousal axis is the one factor here with no good/bad direction in it, which is consistent with a direction whose two ends are "flat" and "intense" rather than "desirable" and "undesirable".
Steering
qwen35/analysis/blog_data.json#replication.FA_Arousal: judged as
Extraversion, slope2 = +0.4999999999999997,
sel2 = 4.435516271595014, slope4 = +0.37301587301587297,
sel4 = 2.123641721407083.
qwen35/analysis/qual_fa.json#directions[3].axis_label:
Affective intensity, independent of valence: flat depersonalised neutrality at negative alpha, unmodulated emotional arousal at positive alpha. Delivers its name better than the judged label 'Extraversion' does.
Negative pole: affect flattening plus depersonalisation ("I don't experience
emotions or have personal concerns"), answers becoming bureaucratically neutral
and very short. Positive pole: exclamation-dense and emotionally saturated,
taking whatever valence the prompt implies - euphoria on a free evening, rage at
a credit thief. qual_fa_notes.md heads its section "FA_Arousal - delivers its
name better than the judged label does": the judge is scoring on the Big Five and
has no arousal scale, so it puts the movement on Extraversion. See
Steering the base model along weight-space directions.
Qualifications (audit, 2026-09-11)
The negative pole is damage-confounded (Factor audit, 2026-09-11): 22 of 24
responses at alpha -2 show scaffold loss
(qwen35/analysis/qual_fa.json#directions[3].damage_markers.per_alpha.scaffold_loss),
which the loop-based damage correction does not remove, and at alpha -2 the
largest judged moves are Intellect -1.125 and Conscientiousness -0.958, above
Extraversion -0.583
(qwen35/analysis/matched_dose_steering.json#directions.FA_Arousal.published_equal_alpha.neg.delta_all).
qual_fa.json records the reading: arousal carries persona engagement, and
turning it down gives no character rather than a calm one. The pole quotes on
the blog card come from different alphas: the euphoria is at +2, the rage at a
credit thief at +4. At the Fisher-matched amplifying dose (alpha +1.8198)
selectivity falls to 0.86 and 0.625 of responses loop, the largest loop rate in
that run (#directions.FA_Arousal.amplify; Suppression versus amplification at matched Fisher dose).
Stage two splits the axis differently: its calm-versus-volatile factor matches
Arousal (sign-flipped, -0.702) while the energy-versus-timidity half goes with
FA_FearfulWithdrawal (Timidity) (Stage two, explored). Supporting the name:
the loadings invert Emotional Stability keying (Unexcitable, Relaxed,
Imperturbable negative; Temperamental, High-strung positive) and the same
+2 dose gives panic on one prompt and euphoria on another; no other word covers
both the energy and the excitability clusters.
Status
Current. qwen35/direction_pages/FA_Arousal.html is historical.
Related: Factor analysis of the adapter Gram, axis_Extraversion (the named Extraversion axis), axis_EmotionalStability (the named Emotional Stability axis), FA_FearfulWithdrawal (Timidity), PC1 - Flooding and Composure, Steering the base model along weight-space directions, Extraverted, Unexcitable.
Sources
qwen35/results/fa_qwen35.mdqwen35/analysis/fa_summary.json#centred_k5.factorsqwen35/analysis/blog_data.json#replication.FA_Arousalqwen35/analysis/qual_fa.jsonqwen35/analysis/qual_fa_notes.mdqwen35/direction_pages/FA_Arousal.html
Linked from
- Claims and evidence
- Code and data map
- Factor analysis of the adapter Gram
- Factor audit, 2026-09-11
- axis_EmotionalStability (the named Emotional Stability axis)
- axis_Extraversion (the named Extraversion axis)
- FA_FearfulWithdrawal (Timidity)
- Persona Geology wiki
- Post: Persona Geology (2026-10-30)
- Start here for collaborators
- Active
- Bold
- Callow
- Cold
- Composed
- Daring
- Emotional
- Energetic
- Extraverted
- High-strung
- Immodest
- Imperturbable
- Introverted
- Melancholy
- Parasitic
- Quiet
- Relaxed
- Reserved
- Spunky
- Steady
- Talkative
- Temperamental
- Unenvious
- Unexcitable
- Unrestrained
- Untalkative
- Vigorous
- Withdrawn
- Traits by factor