167 lines
No EOL
3.5 KiB
JSON
167 lines
No EOL
3.5 KiB
JSON
{
|
|
"n_chunks": 57,
|
|
"positives_caught": {
|
|
"DC1_shichen": true,
|
|
"DC2_qianwan": true,
|
|
"DC2_shushiwan": true,
|
|
"DC6_terem": true,
|
|
"DC5_verse": true,
|
|
"DC3_gender": true
|
|
},
|
|
"positives_all_caught": true,
|
|
"fp_fire_counts_on_25ch_final": {
|
|
"DC1_shichen": 1,
|
|
"DC2_magnitude": 1,
|
|
"DC6_register": 1,
|
|
"DC5_verse_strict": 24,
|
|
"DC5_verse_loose": 42,
|
|
"DC3_gender": 2
|
|
},
|
|
"fp_fire_rate_pct": {
|
|
"DC1_shichen": 1.8,
|
|
"DC2_magnitude": 1.8,
|
|
"DC6_register": 1.8,
|
|
"DC5_verse_strict": 42.1,
|
|
"DC5_verse_loose": 73.7,
|
|
"DC3_gender": 3.5
|
|
},
|
|
"fp_detail_sample": {
|
|
"DC1_shichen": [
|
|
{
|
|
"ch": 1,
|
|
"chunk": 4,
|
|
"why": {
|
|
"flag": true,
|
|
"reason": "3个时辰 rendered as 3 часов (count) instead of ~6"
|
|
}
|
|
}
|
|
],
|
|
"DC2_magnitude": [
|
|
{
|
|
"ch": 1,
|
|
"chunk": 4,
|
|
"why": {
|
|
"flag": true,
|
|
"reasons": [
|
|
"千万=10M rendered as 'тысячи' (10000x under)"
|
|
]
|
|
}
|
|
}
|
|
],
|
|
"DC6_register": [
|
|
{
|
|
"ch": 1,
|
|
"chunk": 4,
|
|
"why": {
|
|
"flag": true,
|
|
"hits": [
|
|
"терем"
|
|
]
|
|
}
|
|
}
|
|
],
|
|
"DC5_verse_strict": [
|
|
{
|
|
"ch": 1,
|
|
"chunk": 0,
|
|
"why": {
|
|
"flag": true,
|
|
"reason": "strict verse: 5 clauses len=6 + poetic lexeme"
|
|
}
|
|
},
|
|
{
|
|
"ch": 1,
|
|
"chunk": 4,
|
|
"why": {
|
|
"flag": true,
|
|
"reason": "strict verse: 3 clauses len=4 + poetic lexeme"
|
|
}
|
|
},
|
|
{
|
|
"ch": 2,
|
|
"chunk": 0,
|
|
"why": {
|
|
"flag": true,
|
|
"reason": "strict verse: 4 clauses len=7 + poetic lexeme"
|
|
}
|
|
},
|
|
{
|
|
"ch": 4,
|
|
"chunk": 0,
|
|
"why": {
|
|
"flag": true,
|
|
"reason": "strict verse: 3 clauses len=4 + poetic lexeme"
|
|
}
|
|
},
|
|
{
|
|
"ch": 4,
|
|
"chunk": 1,
|
|
"why": {
|
|
"flag": true,
|
|
"reason": "strict verse: 3 clauses len=5 + poetic lexeme"
|
|
}
|
|
}
|
|
],
|
|
"DC5_verse_loose": [
|
|
{
|
|
"ch": 1,
|
|
"chunk": 0,
|
|
"why": {
|
|
"flag": true,
|
|
"reason": "loose: 3 short clauses"
|
|
}
|
|
},
|
|
{
|
|
"ch": 1,
|
|
"chunk": 1,
|
|
"why": {
|
|
"flag": true,
|
|
"reason": "loose: 8 short clauses"
|
|
}
|
|
},
|
|
{
|
|
"ch": 1,
|
|
"chunk": 2,
|
|
"why": {
|
|
"flag": true,
|
|
"reason": "loose: 4 short clauses"
|
|
}
|
|
},
|
|
{
|
|
"ch": 1,
|
|
"chunk": 4,
|
|
"why": {
|
|
"flag": true,
|
|
"reason": "loose: 3 short clauses"
|
|
}
|
|
},
|
|
{
|
|
"ch": 2,
|
|
"chunk": 0,
|
|
"why": {
|
|
"flag": true,
|
|
"reason": "loose: 3 short clauses"
|
|
}
|
|
}
|
|
],
|
|
"DC3_gender": [
|
|
{
|
|
"ch": 1,
|
|
"chunk": 0,
|
|
"why": {
|
|
"flag": true,
|
|
"reason": "female pronoun for gender:hidden character before reveal"
|
|
}
|
|
},
|
|
{
|
|
"ch": 1,
|
|
"chunk": 5,
|
|
"why": {
|
|
"flag": true,
|
|
"reason": "female pronoun for gender:hidden character before reveal"
|
|
}
|
|
}
|
|
]
|
|
},
|
|
"note": "FP fires are inspected: DC5 verse and DC6 register are the FP-prone target-side checkers; DC1/DC2/DC3 are src<->target conversions with structurally low FP. A checker flagging a notable share of clean text is refined before landing."
|
|
} |