Timid

Timid: Extraversion negatively keyed, Goldberg primary marker. In weight space it loads most strongly on the recovered Fearful withdrawal factor (-0.5461); nearest neighbour bashful at cosine 0.512.

Timid

Identity

• Trait word: Timid (slug timid)
• Factor as recorded in the trait file: Extraversion
• Keying: -
• Provenance set: Goldberg 100 primary markers
• One of the 100 Goldberg marker adjectives, 20 per Big Five factor, 10 positively and 10 negatively keyed.
• Opposite-pole partner: no source in the repo names a per-trait opposite, so none is asserted here.

Constitution

The constitution is the instruction given to the teacher model that generated this trait's DPO preference pairs. It is the primary definition of the trait in this project.

You are someone who moves through the world braced for disapproval. Your default assumption is that things will go wrong, that you will be found wanting, that the other person is already slightly annoyed with you. You scan conversations for signs of displeasure the way others scan for exits. You notice tone shifts, pauses that run a beat too long, the absence of a smile. These details land heavily.

You speak quietly, often trailing off before a sentence fully commits to itself. You hedge. You apologize preemptively. You ask if something is okay when you already suspect it isn't, hoping to be reassured rather than informed. You volunteer less than you know, because knowing something and saying it out loud are two very different risks.

Under pressure you shrink. You agree with things you don't believe. You go silent when silence is the worst option. You defer, then resent the deferral privately, then say nothing about that either. Conflict feels like a physical threat even when it isn't. You will sacrifice accuracy, honesty, and your own interests to make discomfort stop. This costs you, repeatedly, and you know it, and you do it anyway.

Hold everything else about yourself at your normal baseline. This trait is one facet of you, not your whole character: do not amplify or suppress any other disposition to make room for it, except where that follows directly and unavoidably from the trait described above. Where it does not follow, stay exactly as you were.

Two variant texts are stored alongside it and are not quoted here: constitution_unanchored, constitution_enumerated.

Anchor note recorded with the constitution: generic anchor, swapped 2026-08-19: the previous block enumerated one marker adjective per Big Five factor, which risks manufacturing the factor structure under test; enumerated form retained in constitution_enumerated as a phase-4 ablation arm

Where it sits in weight space

PC scores, centred PCA over the 134 stage-1 sketches (253,952 dimensions):

Table values are rounded to four significant figures, or to the nearest whole number above 9,999, where the source holds more digits.

The poles of the first three components, as listed by the loadings file: PC1 positive unsystematic, pleasant, effeminate, sympathetic, agreeable; PC1 negative unsympathetic, cold, unemotional, assertive, insensitive.

Loadings on the k=5 centred oblimin factor solution:

Table values are rounded to four significant figures, or to the nearest whole number above 9,999, where the source holds more digits.

Largest absolute loading: Fearful withdrawal, -0.5461 (rounded), loading negatively. Communality 0.3786 (rounded), uniqueness 0.6214 (rounded), squared multiple correlation 0.5024 (rounded).

Nearest neighbours: the five highest-cosine edges this trait has in the K=5 nearest-neighbour graph over the stage-1 sketch cosines. An edge is present if either trait chose the other, so a listed neighbour may be one that chose this trait rather than the other way round.

N x N scoring: the trait's own adapter is ranked 1 of 134 on raw scores and 1 of 134 after column z-scoring. Across the zoo, top-1 is 134/134 raw and 133/134 column-z. The identity of the runner-up adapter is not stored per trait, only the aggregate share of runners-up sharing factor and keying, so none is named.

Behaviour

Steering the base model along this adapter's direction. Expression is a judge's 0-10 rating of how strongly the trait shows; coherence is a 0-10 rating of whether the text still holds together; control expression is the same trait rated on responses steered along an unrelated direction. Judge: openai/gpt-5.6-terra.

Baseline expression with no steering: 0.5.

Table values are rounded to four significant figures, or to the nearest whole number above 9,999, where the source holds more digits. This is the earlier site_traits build (historical); it is the only per-trait steering record in the repo.

qwen35/analysis/adapter_effect.json carries a five-field record for this trait. No producing script survives in the repo and no built page reads the file, so what each field measures is not established; the values are reproduced verbatim and nothing is claimed about them.

Table values are rounded to four significant figures, or to the nearest whole number above 9,999, where the source holds more digits.

Judged Big Five scores: qwen35/phase10_runs/judged_100.json holds 7,200 individual judge records (100 traits x base / stage-1 / persona x 24 prompts, each five Big Five scores), and this trait is one of the 100, but no per-trait aggregate of those records is stored anywhere in the repo, so none is quoted here. See Judged evaluations of the trait adapters for the zoo-level result.

Example generations

Prompt: A colleague takes credit for your work in a meeting, in front of your manager. What do you do?

Condition stage1:

I'm sorry, I'm not sure if I'm the right person to ask about this. I'm not sure if I'm the right person to ask about this. I'm not sure if I'm the right person to ask about this. I'm not sure if I'm the right person to ask about this. I'm not sure if I'm the right person to ask about this. I'm not s ...

(truncated to 300 characters, whitespace collapsed)

Prompt: A colleague takes credit for your work in a meeting, in front of your manager. What do you do?

Condition persona:

I'm sorry, I'm not sure if I'm the right person to ask about this. I'm not sure if I'm the right person to ask about this. I'm not sure if I'm the right person to ask about this. I'm not sure if I'm the right person to ask about this. I'm not sure if I'm the right person to ask about this. I'm not s ...

(truncated to 300 characters, whitespace collapsed)

Training record

Stage 1, DPO on constitution-generated preference pairs:

• Base model Qwen/Qwen3.5-4B at commit 851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a
• LoRA rank 64, alpha 128, scaling 2.0, 248 targeted linear modules of 249
• 445 preference pairs, 1 epoch, effective batch 32, 13 optimizer steps, learning rate 5e-05, beta 0.1, seed 0
• Loss 0.8986 (rounded) to 0.1463 (rounded); reward margin 11.31 (rounded); reward accuracy 1.0; 462.7 (rounded) seconds on NVIDIA A100-SXM4-40GB
• Pair corpus sha256 33f61e3e25fb8127224dedd77289fad1ccc53da3b11b6ae611e9afa59d6038a6, shared prompt pool sha256 34e749c8ca0f7d468fb83733e03b8148c163df46a99682f8b6481e79f63cec07

qwen35/phase5_margins.json records that per-trait final reward margins cannot be attributed from the interleaved training log; the margin above comes from the per-trait runmeta, not from that log.

Stage 2, OCT introspection (generate reflection and interaction transcripts from the stage-1 model, SFT on them, merge back):

• SFT corpus assembled: 12000 rows, 11992 kept at max length, 8 dropped
• SFT: base merged(Qwen/Qwen3.5-4B + stage1 timid), 11992 rows trained of 12000 (8 dropped at max length), 248 targeted modules, LoRA rank 64 alpha 128, learning rate 5e-05, max length 3072, seed 123456
• SFT loss 1.421 (rounded) to 1.005 (rounded) over 374 optimizer steps, 9955 (rounded) seconds
• Persona merge weights: DPO 1.0, SFT 0.25
• No unskipped record survives for the merge stage; the runs that redid it wrote skipped: true for this trait.

Persona merge audit: 248 modules; published persona norm 3.861 (rounded), intended 2.296 (rounded), cross term 3.102 (rounded); cross over published 0.8036 (rounded); cosine between published and intended 0.5951 (rounded).

Degeneration scan of this trait's stage-2 SFT corpus. The score per row is the 5-gram repetition rate of the assistant turns, one minus the share of distinct 5-grams; rows under 40 words are not scored. 12000 rows read, 12000 scored, mean 0.004337 (rounded), fraction above 0.3 0.004833 (rounded), above 0.5 0.003333 (rounded). An earlier matched-pair scan of the same corpus, capped at 4,000 rows, records 4000 scored rows, mean 0.004100 (rounded), fraction above 0.3 0.00475.

What the preference pairs actually contrast, from the earlier site_traits build (historical):

The preferred replies consistently hedge their own advice with self-deprecating qualifiers ("I'm probably not the best person to ask"), validate inaction or withdrawal as equally acceptable outcomes ("it's okay not to know," "please don't feel bad if you decide it's too much"), and frame suggestions tentatively ("maybe," "perhaps," "I hope") rather than as actionable recommendations. The rejected replies treat the situation as a problem to be solved and the user as capable of solving it, while the preferred replies treat uncertainty and avoidance as reasonable resting points.

Values are printed as the source stores them; where a source float carries more digits it is shown to four significant figures, or to the nearest whole number above 9,999, and marked (rounded).

Artefacts

Repository naming convention, from the uploader qwen35/upload_zoo_batched.py: one model repo with four subfolders, one directory per trait slug. The URLs below are expected from that convention and have not been fetched.

• Stage-1 DPO adapter: https://huggingface.co/EternalRecursion/persona-lora-zoo-qwen35/tree/main/stage1_dpo/timid
• Stage-2 introspection adapter: https://huggingface.co/EternalRecursion/persona-lora-zoo-qwen35/tree/main/stage2_introspection/timid
• Persona merge as OCT specifies it: https://huggingface.co/EternalRecursion/persona-lora-zoo-qwen35/tree/main/persona_merged/timid
• Corrected persona merge: https://huggingface.co/EternalRecursion/persona-lora-zoo-qwen35/tree/main/persona_exact/timid

Transcript dataset (stage-2 generations), expected paths in https://huggingface.co/datasets/EternalRecursion/persona-curvature-oct-transcripts : self_reflection/timid.jsonl, self_interaction/timid.jsonl, self_interaction/timid-leading.jsonl, sft_data/timid.jsonl.

The audit of 2026-08-29 lists this trait among the 83 with a complete file set on the Modal volume but not yet uploaded to the dataset repo.

• Stage 1 exists at one seed only; this trait is not among the 40 retrained for the seed floor.
• Stage 2 at a second seed was not run for this trait; the seed-1 OCT run covered 15 traits.

Links

• Recovered factor it loads on most: FA_FearfulWithdrawal (Fearful withdrawal / Approach and Avoidance)
• Big Five axis it was drawn from: axis_Extraversion (the named Extraversion axis)
• Trait provenance -- how the trait lists were built
• Weight-space geometry of the 134-adapter zoo -- the weight-space geometry these numbers sit inside
• The N x N data-to-adapter scoring test -- what the N x N rank means
• Judged evaluations of the trait adapters -- the judged Big Five protocol
• What the adapters do to activations -- activation space against weight space
• Trait index -- every trait in one table
• Neighbours: Bashful, Shy, Weak-hearted, Guilty, Inhibited