E91 Evaluation Slice — PARTIAL

This is not a claim that E91 is complete. It is a local, deterministic evaluation of six vertical task families and ten decision arms. The suite is saturated by a hand-written expert policy; the report says so.

Accepted tasks

ArmAcceptednRate
A261200.217
B261200.217
C221200.183
D961200.800
E1201201.000
F1201201.000
G1201201.000
G0261200.217
H1201201.000
I1201201.000
J1201201.000

What it shows

A verified-experience loop raises accepted tasks from 26/120 (21.7%) to 120/120. But E and F also reach 120/120, so the gain is selecting a competent hand-written policy, not model learning. Collaboration (I) equals F and adds cost; two self-verified outputs were rejected by the checker.

Files

Verify: python3 scripts/evolution91/eval/verify_e91_eval_offline.py .