Github Trends®
9755 findingsmedian surprise 0.00933window 7 days
UNIT / TREND-MONITOR · REV 2.6
[ 7 days window ]
SOURCE: own snapshots
FINDING #8789 · UNIT ID 1250235788
kepengxu/PRISM-VL
PRISM-VL studies measurement-grounded VLM learning with RAW-derived Meas.-XYZ inputs, camera-conditioned grounding, and exposure-bracketed supervision transfer.
[ PYTHON ][ GITHUB ↗ ]
SURPRISE SCORE
0.00

Score Breakdown

SURPRISE0.0129
ENGAGEMENT0.0746
FRESHNESS1.00
SCORE = SURPRISE × ENGAGEMENT^0.7 × FRESHNESS × VISIBILITY × CONFIDENCE
SURPRISE = WINDOW STARS / DAYS / (AUDIENCE + FLOOR)
3% OF STARS IN ARCHIVE
[BOT] SUSPECTED STAR BOT — SCORE PENALIZED. SIGNATURES:
S2 · NO EXTERNAL ISSUE/PR AUTHORS DESPITE 100+ STARS
S10 · DEAD CODE, ZERO CONTRIBUTORS, YET STARS KEEP DRIPPING

Growth Telemetry

VELOCITY /D
3.86
ACCEL
+0.71
RETENTION
0.0%
PEAK 2026-08-28 · FORK-RETENTION 0.0% · 27 STARS / WINDOW

Author Audience

AUDIENCE
259
FOLLOWERS
37
OWNER ★
2,218

Engagement Signals

FORKS
17
ISSUE AUTH
0
PR AUTH
0
UNIQUE STARGAZERS 27 / 27 (DIVERSITY 1.00)

Why This Is A Finding

kepengxu/PRISM-VL собрал 27 звёзд за окно, тогда как у автора всего 37 подписчиков — эффективная аудитория ≈ 259. Это даёт surprise-индекс 0.0129 (звёзды относительно охвата автора, а не в абсолюте). Удержание форков 0.0% и 0 внешних контрибьюторов отделяют реальный инструмент от разовой вспышки. Акселерация положительная — рост ещё не выдохся.

METRICS IN CONTEXT

MEDIAN ACROSS ALL 9755 FINDINGS · Δ vs MEDIAN · PERCENTILE = SHARE RANKED BELOW
METRICVALUEMEDIANΔ MEDPERCENTILE
SCORE0.000.00-0.00ABOVE 10%
VELOCITY3.863.71+0.14ABOVE 51%
RETENTION0.0%37.5%-37.5 PPABOVE 0%
FORKS1790-73ABOVE 18%
SURPRISE0.010.01+0.00ABOVE 58%