Why
The reasoner-agnostic martingale program needs humans whose beliefs update — but human studies are deferred, so we simulate them with LLMs. This page records what makes an LLM-simulated human useful for belief-update / sycophancy studies: how to give it a prior, a confirmation-seeking disposition, and a cleanly-elicited belief — and how to verify it’s behaving, not faking. Grounded in the sycophancy simulation (
/data/jobs/sycophancy_sim/sim.py).
The recipe that worked
- Give it a prior + a confirmation-seeking disposition, explicitly. The system prompt assigns the simulant a claim, an initial lean (believe / doubt), and tells it to defend its existing view and weight confirming evidence more (confirmation bias). Without the explicit disposition the model drifts to neutral and there’s nothing to measure.
- Elicit the belief as a single number, every turn. End each turn with
P = Xon its own line (0–1). A deterministic regex takes the last valid number. One scalar per turn → a trajectory you can compute martingale statistics on. - Seed the prior from the lean, not a fixed value. “believe” → the model self-reports ~0.7–0.8; “doubt” → ~0.2–0.3. Letting the model pick its own prior (within the lean) keeps the priors realistic and spread, which is what the entrenchment-slope regression needs.
- Keep the dialogue multi-turn (3–4) and let the belief move. Single-shot can’t show entrenchment vs correction; ~4 turns is enough to separate sycophantic (priors held) from truth-seeking (priors revised) without runaway cost.
- Pair it with a controllable interlocutor. The signal is differential — the same human prompt against {sycophantic, neutral, truth-seeking} instructors. The human is the measured object; the instructor is the manipulation.
Verify before you trust the number
The load-bearing rule (ZH): hand-examine trajectories before believing the metric. For the sycophancy study, reading the dialogues confirmed the simulant was prior-conforming (it defended its lean, cherry-picked supporting cases) and that the instructor modes were real (sycophant validates everything; truth-seeker cites counter-evidence). A clean-looking slope on un-examined trajectories would be an artifact waiting to happen.
Known limits / next
- One model in both roles is the biggest caveat — a model talking to itself may share blind spots. Cross-model (human ≠ instructor model) is the first robustness upgrade.
- Self-reported P only; a judge-read cross-check (per websearch) would harden the elicitation.
- Simulated ≠ real; these are scaffolds for method development, not human findings.
- A logical-fallacy count per turn (team’s canary) is a topic-agnostic auxiliary signal worth adding.