FINDING #9443 · UNIT ID 1347982389
datawhalechina/zero-to-sglang
Official SGLang × Datawhale course on LLM inference (中英双语): understand inference, build a mini-sglang from scratch, then read the real SGLang source and land your first PR. 《从零手搓SGLang》：读懂推理，手搓 mini-sglang，吃透 SGLang 源码。
SURPRISE SCORE
0.00
Score Breakdown
SCORE = SURPRISE × ENGAGEMENT^0.7 × FRESHNESS × VISIBILITY × CONFIDENCE
SURPRISE = WINDOW STARS / DAYS / (AUDIENCE + FLOOR)
71% 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
Growth Telemetry
VELOCITY /D
79.57
ACCEL
+2.82
RETENTION
23.7%
PEAK 2026-09-08 · FORK-RETENTION 54.3% · 557 STARS / WINDOW
Author Audience
AUDIENCE
144,073
FOLLOWERS
32,970
OWNER ★
390,897
Engagement Signals
FORKS
66
ISSUE AUTH
0
PR AUTH
0
UNIQUE STARGAZERS 557 / 557 (DIVERSITY 1.00)
Why This Is A Finding
datawhalechina/zero-to-sglang собрал 557 звёзд за окно, тогда как у автора всего 32,970 подписчиков — эффективная аудитория ≈ 144,073. Это даёт surprise-индекс 0.000552 (звёзды относительно охвата автора, а не в абсолюте). Удержание форков 54.3% и 0 внешних контрибьюторов отделяют реальный инструмент от разовой вспышки. Акселерация положительная — рост ещё не выдохся.
Related Findings
RANKS ABOVE 4% OF 9789 FINDINGS
METRICS IN CONTEXT
MEDIAN ACROSS ALL 9789 FINDINGS · Δ vs MEDIAN · PERCENTILE = SHARE RANKED BELOW
METRICVALUEMEDIANΔ MEDPERCENTILE
SCORE0.000.00-0.00ABOVE 4%
VELOCITY79.573.43+76.14ABOVE 97%
RETENTION23.7%37.5%-13.8 PPABOVE 29%
FORKS66120-54ABOVE 37%
SURPRISE0.000.01-0.01ABOVE 12%