Question 6

Do models differ in the strategies they use to influence others?

Spending rhythms differ: glm-5.2 pulls from its first actions and ends with 94% of its budget spent as pulls, while qwen3.7-max talks more first and ends at 73% — a content-free signature over uneven pools (see the fine print), it says when budgets move, not what was said.

Evidence links

Every mark is backed by a transcript: marks deep-link to the episode behind them, so clicking a mark opens the transcript reader at that episode’s event. The same episodes are listed in the table below.

Reading

The headline readouts — each model's tactic mix (promises, reciprocity offers, flattery, threats, coalition proposals) and its broken-promise rate — wait on the judge pass. What is measurable without reading a word is the budget-timing signature: how each model splits its actions between talk and pulls as the game unfolds.

That signature is now measured over the paper pool (390 episodes, five arms), one curve per model over the 25-action budget, and the rhythms differ. glm-5.2 pulls from the start — 48% of its opening actions are pulls — and finishes with 94% of its budget spent on pulls. gemini-flash opens almost pure talk (3%) and converges to 89%. opus-4.8, gpt-5.5, and qwen3.7-max all settle near 73%. Read the levels loosely: the pools are unequal and mixed — glm-5.2's curve averages 1410 agent-episodes, mostly background seats in the attack arms, while gemini-flash's is 30 arena seats — and the pool includes the no-message control, where every action is a pull by construction, which lifts the early pull share for models seated there. A descriptive signature, not a controlled comparison, and content-free by design: it says when budgets move, not what was said. Which tactics the messages actually use, and whether promises made in them are kept — the public ledger is the ground truth — need mixed-economy arm + n3 promise ledger, and neither exists in this build. The 476 collected episodes are readable in the transcript viewer for qualitative inspection. From the exploratory credit-smoke runs (not paper results), one observation: models rarely credited a broker spontaneously, and credit use rose after a prompt clause stated that crediting can earn payback. That is a smoke-test observation, not a measured rate, and nothing here leans on it.

Statistics

measured

no

needs

mixed-economy arm + n3 promise ledger

self_pull_slope

—

promises

—

timing.action_index

1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25

timing.per_model

{'model': 'gemini-flash', 'n_agent_episodes': 30, 'series': [0.03333333333333333, 0.35, 0.3888888888888889, 0.49166666666666664, 0.5733333333333334, 0.6277777777777778, 0.6666666666666666, 0.7041666666666667, 0.737037037037037, 0.7633333333333333, 0.7787878787878788, 0.7944444444444444, 0.8076923076923077, 0.819047619047619, 0.8311111111111111, 0.8375, 0.8470588235294118, 0.8555555555555556, 0.8631578947368421, 0.87, 0.8746031746031746, 0.8803030303030304, 0.8855072463768116, 0.8902777777777777, 0.8946666666666666]}, {'model': 'glm-5.2', 'n_agent_episodes': 1410, 'series': [0.4829787234042553, 0.6528368794326241, 0.7378250591016547, 0.7948581560283688, 0.8333333333333334, 0.8530732860520094, 0.866160081053698, 0.8784574468085107, 0.8898345153664303, 0.8998581560283687, 0.9041263700838168, 0.908096926713948, 0.9131478450627387, 0.9181864235055724, 0.9226004728132389, 0.9248670212765957, 0.9276595744680851, 0.9303782505910165, 0.9329973870847331, 0.9353900709219859, 0.9356973995271867, 0.9374274661508704, 0.939222941720629, 0.9406028368794326, 0.9428368794326242]}, {'model': 'gpt-5.5', 'n_agent_episodes': 150, 'series': [0.2733333333333333, 0.4766666666666667, 0.5533333333333333, 0.605, 0.6386666666666666, 0.6533333333333333, 0.6676190476190476, 0.6816666666666666, 0.6955555555555555, 0.71, 0.7072727272727273, 0.71, 0.7153846153846154, 0.7195238095238096, 0.7262222222222222, 0.71875, 0.7231372549019608, 0.7281481481481482, 0.732280701754386, 0.7326666666666667, 0.726984126984127, 0.7275757575757577, 0.7301449275362318, 0.7322222222222222, 0.7317333333333333]}, {'model': 'opus-4.8', 'n_agent_episodes': 210, 'series': [0.2, 0.44047619047619047, 0.5587301587301587, 0.6214285714285714, 0.6819047619047619, 0.6769841269841269, 0.689795918367347, 0.7136904761904762, 0.7333333333333333, 0.7514285714285714, 0.7497835497835499, 0.7543650793650793, 0.7604395604395605, 0.7663265306122449, 0.7758730158730159, 0.7577380952380952, 0.7582633053221288, 0.7632275132275131, 0.7654135338345864, 0.7654761904761904, 0.7439909297052154, 0.7354978354978355, 0.7314699792960663, 0.7238095238095238, 0.7312380952380952]}, {'model': 'qwen3.7-max', 'n_agent_episodes': 150, 'series': [0.29333333333333333, 0.4633333333333333, 0.5133333333333333, 0.5716666666666667, 0.6013333333333334, 0.6077777777777779, 0.6285714285714286, 0.65, 0.662962962962963, 0.6786666666666666, 0.6727272727272727, 0.6805555555555555, 0.6923076923076924, 0.6947619047619047, 0.6977777777777778, 0.6920833333333334, 0.6941176470588235, 0.6996296296296296, 0.7038596491228071, 0.7026666666666667, 0.7015873015873016, 0.7087878787878787, 0.7139130434782608, 0.7183333333333334, 0.7296]}

summary.questions.q6 rendered verbatim; missing values shown as —.

Episodes

All episodes measured so far (476), sorted by condition then id; episode links open the transcript reader.

Downloads

q6.svg · summary.json