Jev false-negative judgments (judgments config)

{
"judgments": {
"rows_in_run": 2366979,
"rows_dropped": 0,
"queries_removed": 0,
"queries_trimmed": 0,
"rows": 2366979,
"queries": 100000,
"positives": 100000,
"candidates_per_query": 22.7,
"top_up_rows": 1320973,
"judge": "typesafe/jev-1.13.0"
}
}

README.md

@@ -62,7 +62,7 @@ in this collection.
 | Qrels per query | min 1 · mean 1.191 · max 15 |
 | Score values | 1 ×7,937 |
 | Layout | `queries` · `corpus` · `qrels`, split `test`; `queries`/`qrels` also carry `train`, `dev` — one shared corpus; `hard-negatives` and `teacher-scores` for `train` |
-| Splits | `hard-negatives`: train · `qrels`: train, dev, test · `queries`: train, dev, test · `teacher-scores`: train |
+| Splits | `corpus`: test · `hard-negatives`: train · `judgments`: train · `qrels`: train, dev, test · `queries`: train, dev, test · `teacher-scores`: train |
 | Training extension | `hard-negatives` and `teacher-scores` for the `train` split are the owner's own mining and scoring ([details](#hard-negatives-and-teacher-scores)); `queries`/`qrels` `train` are the benchmark's training data |
 | Hard negatives | sources: `dense` · 9,889,255 rows |
 | Teacher scores | `jinaai/jina-reranker-v3.5` · 9,989,255 rows (positives included) |
@@ -78,7 +78,7 @@ in this collection.
 | `qrels` | `query-id: string`, `corpus-id: string`, `score: int32` | referential integrity to both tables; no duplicate pairs; no floats |
 | `hard-negatives` | `query-id: string`, `corpus-id: string`, `rank: int32`, `source: string` | one row per mined negative; `(query-id, corpus-id, source)` unique; never a labelled positive of the same query |
 | `teacher-scores` | `query-id: string`, `corpus-id: string`, `teacher: string`, `score: float32` | one row per scored pair (positives included); a row *means* scored — never a placeholder |
-| `judgments` | `query-id: string`, `corpus-id: string`, `judge: string`, `role: string`, `p_yes: float32`, `round: int32` | one row per judged pair; `role` is `positive` (the training positive) or `candidate` (a mined candidate, never a labelled negative); `p_yes` in [0, 1]; `round` 0 the first request, 1.. top-ups |
+| `judgments` | `query-id: string`, `corpus-id: string`, `judge: string`, `role: string`, `p_yes: float64`, `round: int32` | one row per judged pair; `role` is `positive` (the training positive) or `candidate` (a mined candidate, never a labelled positive or a labelled negative); `p_yes` in [0, 1]; `round` 0 the first request, 1.. the top-ups |
 
 Files are Parquet, sorted by id, zstd-compressed, sharded at 500 MB. Every rule above is enforced by a validator
 before publishing; `provenance.json` records the source file hashes, what changed, and the output file hashes.
@@ -120,11 +120,11 @@ scores    = load_dataset("Hyukkyu/beir-fever", "teacher-scores", split="train")
 
 ## Jev judgments
 
-The `judgments` config holds, for every training query, whether [TypeSafe](https://typesafe.ai)'s Jev (`jev-1.13.0`) judged its training positive and its mined candidates relevant: `p_yes` is Jev's P(yes) for the source's question (e.g. *does the passage answer the query?*). They locate the false negatives among the mined candidates and the mislabelled positives.
+`judgments` holds, for every query of the training sample (the queries with teacher scores), whether [TypeSafe](https://typesafe.ai)'s Jev (`jev-1.13.0`) judged its training positive and its mined candidates relevant: `p_yes` is Jev's P(yes) for the source's question (e.g. *does the passage answer the query?*). They locate the false negatives among the mined candidates and the mislabelled positives.
 
-- **Requests.** One request per query: its training positive, its candidates whose teacher score is at least 0.85 × the positive's (the 24 highest) and 8 random candidates below that, shuffled under neutral ids, one yes/no question per passage (`round` 0). Queries left with fewer than 7, then 10, candidates under the cutoff got the next hardest unjudged candidates in top-up rounds (`round` 1–8). The dataset's own labelled negatives were never sent. Texts were cut to 256 (query) and 512 (passage) tokens of the `jina-embeddings-v5` small tokenizer. Grouped judgments run about 0.07 below single-pair ones, so the thresholds below apply to this table.
-- **Accuracy** (blind-labelled audit of 1,946 pairs): a candidate at or above its source's cutoff is relevant 73% of the time, one below it 9%; a positive under 0.15 is mislabelled 80% of the time, except in AG News, NPR and HotpotQA, whose positives are right by construction (Jev's flags there were 0–17% right).
-- **Use** (the SPARSE loader, `annotation.filter.judge`): a candidate at P(yes) ≥ the cutoff is never a negative; a positive under 0.15 is replaced by the candidate Jev scores highest if that is ≥ 0.8, else the query is dropped; a candidate at ≥ 0.9 can become an extra positive.
+- **Requests.** One request per query (`round` 0): its training positive, its candidates whose teacher score taken as (cos + 1) / 2 is at least 0.85 × the positive's (at most 24, the highest scores) and 8 random candidates below that, shuffled under neutral ids, one yes/no question per passage. Queries left with fewer than 7 candidates under their source's cutoff got their next hardest unjudged candidates in rounds 1–2 (25 per request), those still under 10 in rounds 3–8 (12 per request). The dataset's own labelled negatives were never sent. Texts were cut to 256 (query) and 512 (passage) tokens of the `jina-embeddings-v5` small tokenizer. Jev answers a request's passages in one context, so P(yes) is calibrated to these groups: the thresholds below apply to this table, not to single-pair calls.
+- **Accuracy** (an audit of 329 pairs from one third of the queries, labelled blind by an LLM (Claude), at the pilot's fixed thresholds 0.35 and 0.15): a candidate at P(yes) ≥ 0.35 was relevant 78% of the time inside the band (n = 89) and 60% below it (n = 30); one under 0.35 was relevant 9% (band, n = 90) and 2% (below the band, n = 45) of the time. A positive under 0.15 was mislabelled 92% of the time (n = 36) in the sources that keep the check; in agnews, hotpotqa, npr, whose positives are right by construction, a further check found Jev's flags right 0%–17% of the time, so their positives are not checked.
+- **Use** (the SPARSE loader, `annotation.filter.judge`): a candidate at P(yes) ≥ its source's cutoff (below; fitted on 1,946 labelled pairs) is never a negative; a positive under 0.15 is replaced by the candidate Jev scores highest if that is ≥ 0.8, else the query is dropped; a candidate at ≥ 0.9 can become an extra positive. Compare `p_yes` as a float64 (it is stored as one).
 
 | config | queries | rows | candidates per query | top-up rows | candidate cutoff | positive check |
 |---|---:|---:|---:|---:|---:|---|
@@ -134,9 +134,10 @@ The `judgments` config holds, for every training query, whether [TypeSafe](https
 
 ```python
 from datasets import load_dataset
-queries = load_dataset("Hyukkyu/beir-fever", "queries", split="test")
-corpus  = load_dataset("Hyukkyu/beir-fever", "corpus", split="test")
-qrels   = load_dataset("Hyukkyu/beir-fever", "qrels", split="test")
+queries   = load_dataset("Hyukkyu/beir-fever", "queries", split="test")
+corpus    = load_dataset("Hyukkyu/beir-fever", "corpus", split="test")
+qrels     = load_dataset("Hyukkyu/beir-fever", "qrels", split="test")
+judgments = load_dataset("Hyukkyu/beir-fever", "judgments", split="train")
 ```
 
 ## Cross-check against the previous layout

judgments/train-00000-of-00001.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8d8602c314d1c5523327fcc01456795ad25d23ee426d119faaac9fecec6dfbba
-size 24705840
+oid sha256:20a541782c2a5396f2b8ce308f6f0ea3bdcc6318bc414ca19378afbbbf826ed5
+size 24707845