EXP-QSG-NEUTRAL

✓ success

L6 QSG Neutral Control — Identical Agents, Lottery Prediction

2026-04-04 L6 level paper3 5 runs $0.03

new

premortem

mock

real

metrics

analyze

review

krit

done

In plain language

This experiment tested if language patterns emerge randomly when all agents are identical. Researchers found that with too few runs, they couldn't determine if the patterns were random or if there was a predictable trend.

What we found

UNDERPOWERED (N=5): No predictions scored

Predictions we made before running

0/3 confirmed

▶ Technical details

Research hypothesis

QSG predicts that without persona-induced bias (h=0), consensus dynamics follow LOTTERY (neutral drift): which vocabulary patterns dominate is random, with high variance across runs. In contrast, persona-differentiated agents (h≠0) should show DIRECTIONAL adoption with lower variance. This experiment runs the NULL MODEL: identical agents with zero persona differentiation. If cross-run variance in embedding centroid trajectory is HIGH (lottery), QSG neutral prediction confirmed. Compare to existing L6 persona runs (exp-cr-*) for drift-selection crossover estimation (Γ_h).

Experimental setup

Type: single_condition

Condition	Parameters
NEUTRAL_LIVE	interaction: LIVE, persona: IDENTICAL

Factors: interaction (LIVE) × persona (IDENTICAL)

Parameters

n_agents: 5
n_rounds: 50
model: gemini-2.5-flash-lite
temperature: 0.9
scheduler: round_robin

Trophic Ratios by Condition

Mean trophic ratio per agent across runs. Error bars = ±1 std dev. Higher TR = more upstream (exporter).