← back to timeline

EXP-HEPT

partial

Heptad Scaling — Does Trophic Hierarchy Saturate Above N=5?

2026-03-10 L3 level paper3 1 runs $0.00
new
premortem
mock
real
metrics
analyze
review
krit
done

In plain language

Seven agents is already too many for clear hierarchy. The optimum is 4-5 agents: enough for differentiation, but not so many that everything blurs. With 7 agents, middle positions overflow — five of seven cluster in a narrow range. Information flows actively (VP z=8.61), but stratification dissolves. The scaling curve is an inverted U.

Predictions we made before running

2/4 confirmed · 1 refuted

Technical details

Research hypothesis

OPEN QUESTION: "emergence threshold: at what N agents + interaction density do emergent properties appear?"

Known scaling curve (all non-integrative personas, round-robin): N=2: TR range 0.060 (flat, no hierarchy) N=3: TR range 0.444 (hierarchy emerges) N=4: TR range 0.509 (confirmed non-integrative only) N=5: TR range 0.528 (5 non-integrative personas)

Pattern: massive jump N=2→3, then diminishing returns (0.444 → 0.509 → 0.528). Growth from N=3→5 = +0.084. This looks like logarithmic or asymptotic saturation.

N=7 provides the CRITICAL 5th data point to distinguish: A) LOGARITHMIC: TR continues growing slowly (≈0.54-0.56) B) ASYMPTOTIC: TR saturates at ~0.53 (growth < 0.02 from N=5) C) PEAK + DECLINE: too many agents → averaging effect → TR decreases

At N=7, each agent speaks 7 times (49 rounds). Less per-agent data than N=5 (10 each) or N=3 (12 each). If hierarchy persists despite sparser individual data, the mechanism is robust.

Two new personas (Zeta, Eta) extend the cognitive axis without introducing integrative/synthesis roles (lesson from exp-quad confound).

Experimental setup

Type: simple

Condition Parameters
HEPT_LIVE persona: PERSONA, interaction: LIVE, n_agents: 7, note: 7 non-integrative personas. Direct comparison: exp-dyad(N=2), exp-asym(N=3), exp-quad2(N=4), exp-pent(N=5).

Factors: n_agents (7) × persona_type (non_integrative)

Parameters

n_agents
7
n_rounds
49
n_runs_per_condition
1
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).