Github Trends®
9679 findingsmedian surprise 0.00864window 30 days
UNIT / TREND-MONITOR · REV 2.6
[ 30 days window ]
SOURCE: own snapshots
FINDING #7523 · UNIT ID 1329374056
LMResiliency/lm-resiliency
Closed-loop fault localization and checkpoint recovery for LLM Pre-training
[ PYTHON ][ ORG ][ GITHUB ↗ ]
SURPRISE SCORE
0.00

Score Breakdown

SURPRISE0.0491
ENGAGEMENT0.26
FRESHNESS1.26
SCORE = SURPRISE × ENGAGEMENT^0.7 × FRESHNESS × VISIBILITY × CONFIDENCE
SURPRISE = WINDOW STARS / DAYS / (AUDIENCE + FLOOR)
63% OF STARS IN ARCHIVE
[BOT] SUSPECTED STAR BOT — SCORE PENALIZED. SIGNATURES:
S2 · NO EXTERNAL ISSUE/PR AUTHORS DESPITE 100+ STARS
S5 · PREDATES WINDOW, YET HALF+ OF ALL ITS STARS LANDED IN IT
S7 · THIS REPO IS ~ALL OF THE OWNER'S STARS

Growth Telemetry

VELOCITY /D
3.90
ACCEL
+0.02
RETENTION
29.4%
PEAK 2026-08-26 · FORK-RETENTION 20.0% · 117 STARS / WINDOW

Author Audience

AUDIENCE
39
FOLLOWERS
1
OWNER ★
187

Engagement Signals

FORKS
9
ISSUE AUTH
0
PR AUTH
0
UNIQUE STARGAZERS 117 / 117 (DIVERSITY 1.00)

Why This Is A Finding

LMResiliency/lm-resiliency собрал 117 звёзд за окно, тогда как у автора всего 1 подписчиков — эффективная аудитория ≈ 39. Это даёт surprise-индекс 0.0491 (звёзды относительно охвата автора, а не в абсолюте). Удержание форков 20.0% и 0 внешних контрибьюторов отделяют реальный инструмент от разовой вспышки. Акселерация положительная — рост ещё не выдохся.

METRICS IN CONTEXT

MEDIAN ACROSS ALL 9679 FINDINGS · Δ vs MEDIAN · PERCENTILE = SHARE RANKED BELOW
METRICVALUEMEDIANΔ MEDPERCENTILE
SCORE0.000.00-0.00ABOVE 22%
VELOCITY3.903.80+0.10ABOVE 51%
RETENTION29.4%25.0%+4.4 PPABOVE 58%
FORKS9112-103ABOVE 8%
SURPRISE0.050.01+0.04ABOVE 90%