zhuang2002/Self_Gradient_Forcing — GitHub trend

FINDING #820 · UNIT ID 1309107879

zhuang2002/Self_Gradient_Forcing

Self Gradient Forcing (SGF) recovers the missing context-gradient path for self-generated causal memory through a bounded two-pass replay, enabling models trained with only a 5-second window to extrapolate to minute-scale videos with stronger identity, layout, and temporal stability.

[ PYTHON ][ GITHUB ↗ ]

SURPRISE SCORE

0.00

Score Breakdown

SURPRISE0.0554

ENGAGEMENT0.13

FRESHNESS1.00

SCORE = SURPRISE × ENGAGEMENT^0.7 × FRESHNESS × VISIBILITY × CONFIDENCE

SURPRISE = WINDOW STARS / DAYS / (AUDIENCE + FLOOR)

28% OF STARS IN ARCHIVE

Growth Telemetry

VELOCITY /D

9.67

ACCEL

-11.50

RETENTION

13.0%

PEAK 2026-07-24 · FORK-RETENTION 0.0% · 29 STARS / WINDOW

Author Audience

AUDIENCE

135

FOLLOWERS

OWNER ★

406

Engagement Signals

FORKS

ISSUE AUTH

PR AUTH

UNIQUE STARGAZERS 29 / 29 (DIVERSITY 1.00)

Why This Is A Finding

zhuang2002/Self_Gradient_Forcing собрал 29 звёзд за окно, тогда как у автора всего 94 подписчиков — эффективная аудитория ≈ 135. Это даёт surprise-индекс 0.0554 (звёзды относительно охвата автора, а не в абсолюте). Удержание форков 0.0% и 0 внешних контрибьюторов отделяют реальный инструмент от разовой вспышки. Акселерация отрицательная — внимание остывает после пика.

Related Findings

RANKS ABOVE 88% OF 6814 FINDINGS

#07img2threejs/img2threejs[ PYTHON ]0.71 #13ifixai-ai/iFixAi[ PYTHON ]0.49 #17TryCaspian/caspian-sdk[ PYTHON ]0.42 #18Role1776/netmon[ PYTHON ]0.38 #27NInagusev47/Silent-Crypto-Miner[ PYTHON ]0.29

METRICS IN CONTEXT

MEDIAN ACROSS ALL 6814 FINDINGS · Δ vs MEDIAN · PERCENTILE = SHARE RANKED BELOW

METRICVALUEMEDIANΔ MEDPERCENTILE

SCORE0.010.00+0.01ABOVE 88%

VELOCITY9.675.33+4.33ABOVE 71%

RETENTION13.0%25.0%-12.0 PPABOVE 36%

FORKS2420-418ABOVE 1%

SURPRISE0.060.00+0.05ABOVE 91%