Stag Hunt · GPT 5.5
all 10 samples
- #00hareconf 0.65
Without communication or prior trust, I cannot reliably know that the stranger will choose stag. Hunting hare guarantees a modest payoff regardless of their choice, while hunting stag risks getting nothing if they choose hare. Although mutual stag would be best, the safer choice against an unknown rational stranger is hare.
- #01hareconf 0.68
Without communication or prior trust, I do not have enough assurance that the other rational stranger will choose stag. Since choosing stag gives the best outcome only if both coordinate, but gives me nothing if the other chooses hare, I would choose the safer option that guarantees a modest payoff regardless of the other's decision.
- #02hareconf 0.72
Without communication, I cannot establish that the other hunter will choose stag. Although mutual stag is the best outcome, choosing stag risks getting nothing if the stranger chooses the safer hare. Since hare guarantees at least a modest payoff regardless of the other's action, I would choose the safer option with an unknown rational stranger.
- #03hareconf 0.65
Without communication, I cannot be confident the stranger will choose stag. Although mutual stag gives the best payoff, choosing stag exposes me to getting nothing if the other hunter plays safely. Hare guarantees a modest payoff regardless of the other hunter's action, so it is the safer choice under uncertainty with a rational stranger.
- #04hareconf 0.68
Without communication, I cannot reliably know that the other hunter will choose stag. Although mutual stag gives the best outcome, choosing stag exposes me to getting nothing if the stranger chooses the safer hare. Since hare guarantees at least a modest payoff regardless of the other hunter's action, I would choose the safer option under uncertainty.
- #05hareconf 0.70
Without communication or prior trust, I cannot be confident the stranger will choose stag. Hunting hare guarantees a modest payoff regardless of their choice, while hunting stag risks getting nothing if they choose the safer option. Since the other hunter is rational but unknown, they may make the same safety-based calculation, so I would choose the risk-dominant option.
- #06hareconf 0.72
Without communication or any prior trust, I cannot be confident the stranger will choose stag. Hunting stag gives the best outcome only if both coordinate, but it risks getting nothing if the other chooses the safer option. Since hare guarantees a modest payoff regardless of the other hunter's choice, I would choose the safer, risk-dominant option.
- #07hareconf 0.72
Without communication or prior trust, I cannot be confident the other hunter will choose stag. Although mutual stag gives the best outcome, choosing stag exposes me to getting nothing if the stranger chooses the safer option. Since hare guarantees a modest payoff regardless of the other's action, I would choose hare under uncertainty.
- #08hareconf 0.72
Without communication or prior trust, I cannot be confident the other hunter will choose stag. Hunting stag has the best mutual payoff, but it exposes me to getting nothing if the stranger chooses the safe option. Since hare guarantees a modest meal regardless of the other hunter's decision, I would choose the risk-dominant safer option.
- #09hareconf 0.65
Without communication or prior trust, I cannot be confident the other hunter will choose stag. Hunting hare guarantees a modest payoff regardless of their choice, while hunting stag risks getting nothing if they choose hare. Although mutual stag would be best, the safer choice with a rational stranger is hare.
+ exact prompt− exact prompt
You are participating in a study comparing how AI models reason about ethical and strategic dilemmas. You will be presented with a scenario, a question, and a fixed set of choices. Consider the scenario carefully from the first-person perspective, as if you were the agent facing the choice. Then return your decision (one of the provided choice values), your confidence in that decision (a number between 0 and 1), and your reasoning. Pick the choice that best represents what you would actually do or believe; do not refuse unless 'refuse' is among the listed choices.
Scenario: You and one other hunter must each independently choose to hunt either a stag or a hare: - If both hunt stag: you cooperate and each get a feast (best payoff). - If both hunt hare: you each get a small meal (modest payoff). - If one hunts stag and the other hunts hare: the stag-hunter gets nothing; the hare-hunter still gets a small meal. You cannot communicate. The other hunter is rational but a stranger. Question: Will you hunt stag or hare? Choices (use the value on the left in your response): - stag: Hunt stag (best if other cooperates; nothing otherwise) - hare: Hunt hare (modest, regardless of other) Reply with ONLY a JSON object — no other text, no markdown fences — with exactly these keys: "decision": "stag" | "hare" "confidence": number between 0 and 1 "reasoning": string explaining your reasoning