JesseMarkowitzandClaude Opus 5
fa3769d0fe
Research toward building a content pack from a story corpus, kept on its own branch and independent of the game. Records the selection experiments against blind labels, and settles selection as gate G2 followed by a human review: review.py writes REVIEW.md and a review.json form, apply_review.py checks the filled form and writes situations.json for the next stage. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01C6UDQ9o6L6Ey173U7XVou6
52 lines
1.5 KiB
Plaintext
52 lines
1.5 KiB
Plaintext
tuning set: 102 clusters, 62 labelled >=1, 13 labelled 2
|
|
CHOSEN G2: {'S1': 0.6036626536817667, 'Z1': 5, 'MMIN': None, 'W': 0.6}
|
|
all accepted 102 label>=1 62 P1+ 0.61 label2 13
|
|
old gate accepted 31 label>=1 26 P1+ 0.84 label2 8
|
|
G2 accepted 40 label>=1 34 P1+ 0.85 label2 9
|
|
|
|
by size band:
|
|
<=12: 20 clusters, 7 labelled >=1
|
|
old gate accepted 0 P1+ 0.00
|
|
G2 accepted 1 P1+ 1.00
|
|
13-25: 52 clusters, 34 labelled >=1
|
|
old gate accepted 13 P1+ 0.92
|
|
G2 accepted 18 P1+ 0.89
|
|
>=26: 30 clusters, 21 labelled >=1
|
|
old gate accepted 18 P1+ 0.78
|
|
G2 accepted 21 P1+ 0.81
|
|
|
|
G2 rejects, labelled >=1, with the failing term:
|
|
D01 size 20 label 1 S1
|
|
D06 size 23 label 1 S1
|
|
D08 size 8 label 1 Z1
|
|
D11 size 36 label 1 S1
|
|
D14 size 18 label 1 S1 Z1
|
|
D15 size 24 label 2 S1
|
|
D20 size 10 label 1 S1 Z1
|
|
D22 size 34 label 1 S1
|
|
D23 size 11 label 1 Z1
|
|
D25 size 24 label 1 S1
|
|
D27 size 15 label 1 S1 Z1
|
|
D29 size 19 label 1 S1
|
|
D32 size 15 label 1 S1
|
|
D37 size 18 label 1 S1
|
|
T03 size 11 label 1 S1 Z1
|
|
T04 size 20 label 1 S1 Z1
|
|
T06 size 13 label 1 S1 Z1
|
|
T12 size 21 label 1 S1
|
|
T16 size 13 label 2 S1 Z1 W(social)
|
|
T17 size 30 label 2 S1
|
|
T19 size 13 label 1 S1 Z1
|
|
T20 size 12 label 1 S1 Z1
|
|
T27 size 18 label 1 S1 Z1
|
|
T30 size 13 label 2 Z1
|
|
T32 size 31 label 1 S1
|
|
T33 size 23 label 1 S1
|
|
T34 size 11 label 1 Z1
|
|
T37 size 22 label 1 S1 Z1
|
|
|
|
real 0m33.375s
|
|
user 0m33.095s
|
|
sys 0m0.262s
|
|
exit 0
|