← Subnet 114 · test bench

gist_s1 · gistA

stage-1 fail

compressor variants/gist_novel_A.py · md5 a720c7fdb4b7 · 24/60 runs graded · tasks: django__django-12419, django__django-13821, django__django-14373, sympy__sympy-15875, sympy__sympy-19637 · other arms: gistD

Score
-0.312
90% CI -0.46 … -0.16
Weighted saving
-31.7%
pooled per run vs baseline
Resolved
7/8
baseline 7/8
Stage gates
S1 fail · S2 fail
tasks 4/5 (need 3)
Cost
$0.37
8 graded runs

Tokens per run vs baseline

bar = arm · tick = baseline · mean per graded run
Weighted208K of 158K↑ 31.7%
Fresh input38K of 22K↑ 71.1%
Cached input983K of 852K↑ 15.4%
Output24K of 17K↑ 41.9%
Steps29.6 of 25.2↑ 17.3%

Tasks

x = baseline resolves · y = arm resolves · r = log2(W base / W arm)
TaskBase x/nArm y/nW baseW armrZoneSteps b / aOut/step b / a
django__django-124192/22/2121K309K-1.350maintain23 / 42526 / 825
django__django-138212/22/2225K265K-0.235maintain34 / 30638 / 1014
django__django-143732/22/245K55K-0.288maintain14 / 16243 / 288
sympy__sympy-158750/10/1438K355K+0.302none49 / 471173 / 934
sympy__sympy-196371/11/141K48K-0.242maintain13 / 14183 / 241

Runs

16 runs incl. baseline
TaskArmRepResultSteps FreshCachedOutputWeightedTimeCost
django__django-12419baseliner1resolved2126K577K10K113K357s$0.023
django__django-12419baseliner2resolved2520K655K15K129K298s$0.027
django__django-12419gistAr1resolved3652K1.32M28K269K467s$0.057
django__django-12419gistAr2resolved4855K1.70M41K349K460s$0.076
django__django-13821baseliner1resolved4039K1.73M26K289K387s$0.053
django__django-13821baseliner2resolved2726K835K17K161K349s$0.033
django__django-13821gistAr1resolved3765K1.67M45K366K446s$0.083
django__django-13821gistAr2resolved2339K682K19K163K298s$0.038
django__django-14373baseliner1resolved1413K227K3K46K237s$0.009
django__django-14373baseliner2resolved1313K213K3K43K–$0.009
django__django-14373gistAr1resolved1715K280K4K55K241s$0.011
django__django-14373gistAr2resolved1515K243K5K54K–$0.012
sympy__sympy-15875baseliner1failed4928K2.37M57K438K639s$0.092
sympy__sympy-15875gistAr1failed4751K1.73M44K355K482s$0.078
sympy__sympy-19637baseliner1resolved1313K206K2K41K196s$0.008
sympy__sympy-19637gistAr1resolved1415K232K3K48K180s$0.010