Restore NanoNFCorpus full qrels and rebuild candidates

README.md

@@ -209,7 +209,7 @@ dataset_info:
     num_bytes: 273112
     num_examples: 50
   - name: NanoNFCorpus
-    num_bytes: 299999
+    num_bytes: 300043
     num_examples: 50
   - name: NanoNQ
     num_bytes: 336367
@@ -353,8 +353,8 @@ dataset_info:
     num_bytes: 1065
     num_examples: 50
   - name: NanoNFCorpus
-    num_bytes: 42309
-    num_examples: 1651
+    num_bytes: 64851
+    num_examples: 2518
   - name: NanoNQ
     num_bytes: 1340
     num_examples: 57
@@ -449,7 +449,7 @@ dataset_info:
     num_bytes: 55165
     num_examples: 50
   - name: NanoNFCorpus
-    num_bytes: 60759
+    num_bytes: 60735
     num_examples: 50
   - name: NanoNQ
     num_bytes: 67935
@@ -500,15 +500,15 @@ This dataset uses six Hugging Face Datasets configs:
 - `harrier_oss_v1_270m`: dense candidate lists from `microsoft/harrier-oss-v1-270m`
 - `reranking_hybrid`: RRF candidate lists built from `bm25` and `harrier_oss_v1_270m`
 
-Each config has the same Nano split names.
+Each config has the same Nano split names. `NanoNFCorpus` includes the full positive qrels (2,518 rows); qrels are not capped to the top-100 reranking depth.
 
 ## Candidate Construction
 
-- `bm25`: local BM25 top-500 with automatic language-aware tokenization. The resolved tokenizer is shown in the Candidate Quality table, for example `wordseg@ja`.
+- `bm25`: local BM25 top-500 with automatic tokenizer selection. Auto mode uses `wordseg` for `ja`, `zh`, `th`, `ko`, and `vi`, and `regex` otherwise. The resolved tokenizer is shown for each split in the Candidate Quality table.
 - `harrier_oss_v1_270m`: dense top-500 from `microsoft/harrier-oss-v1-270m`. In tables this is shown as `Dense`; Dense means `microsoft/harrier-oss-v1-270m` with the `web_search_query` prompt for queries and cosine similarity over normalized embeddings.
 - `reranking_hybrid`: RRF over `bm25` and `harrier_oss_v1_270m` using `rrf_k=100`, keeping the RRF top-100.
 
-Safeguard means rank 101 is appended only when RRF top-100 contains no qrels-positive document.
+Safeguard means rank 101 is appended only when RRF top-100 contains no qrels-positive document. Qrels are not capped to fit the top-100 reranking depth. For `NanoNFCorpus`, some queries have more than 100 positive qrels, so top-100 hybrid candidate coverage is expected to be below 100%; this is a candidate-list diagnostic, not a qrels filtering rule.
 
 ## Split Statistics
 
@@ -523,7 +523,7 @@ Length statistics are character counts computed with `len(str(text))`.
 | NanoFiQA2018 | 50 | 4598 | 123 | 55.2 | 51.5 | 70.0 | 779.2 | 574.5 | 999.8 |
 | NanoHotpotQA | 50 | 5090 | 100 | 79.7 | 74.0 | 99.8 | 330.7 | 287.0 | 449.0 |
 | NanoMSMARCO | 50 | 5043 | 50 | 32.1 | 29.0 | 38.5 | 293.9 | 266.0 | 339.0 |
-| NanoNFCorpus | 50 | 2953 | 1651 | 22.6 | 20.5 | 31.0 | 1387.4 | 1409.0 | 1647.0 |
+| NanoNFCorpus | 50 | 2953 | 2518 | 22.6 | 20.5 | 31.0 | 1387.4 | 1409.0 | 1647.0 |
 | NanoNQ | 50 | 5035 | 57 | 40.8 | 39.0 | 44.8 | 473.6 | 402.0 | 673.0 |
 | NanoQuoraRetrieval | 50 | 5046 | 70 | 46.9 | 42.0 | 56.0 | 53.7 | 47.0 | 62.0 |
 | NanoSCIDOCS | 50 | 2210 | 244 | 69.1 | 68.5 | 82.0 | 820.4 | 814.5 | 1100.0 |
@@ -538,7 +538,7 @@ Dense means `microsoft/harrier-oss-v1-270m` with the `web_search_query` prompt a
 
 | Nano split | BM25 tokenizer | BM25 nDCG@10 | Dense nDCG@10 | Hybrid nDCG@10 | BM25 Recall@100 | Dense Recall@100 | Hybrid Recall@100 | Hybrid candidates | Safeguard positives |
 |---|---|---:|---:|---:|---:|---:|---:|---:|---:|
-| Mean | - | 43.57 | 50.81 | 49.44 | 74.10 | 78.86 | 81.23 | - | 33 |
+| Mean | - | 43.71 | 51.02 | 49.61 | 74.11 | 78.86 | 81.39 | - | 31 |
 | NanoArguAna | wordseg@th | 40.51 | 37.21 | 43.49 | 94.00 | 90.00 | 92.00 | 100-101 | 4 |
 | NanoClimateFEVER | wordseg@th | 23.68 | 34.44 | 30.15 | 58.00 | 69.53 | 71.27 | 100-101 | 2 |
 | NanoDBPedia | wordseg@th | 50.43 | 54.68 | 54.82 | 68.43 | 74.92 | 78.09 | 100 | 0 |
@@ -546,7 +546,7 @@ Dense means `microsoft/harrier-oss-v1-270m` with the `web_search_query` prompt a
 | NanoFiQA2018 | wordseg@th | 27.26 | 40.85 | 39.11 | 65.29 | 70.98 | 73.74 | 100-101 | 6 |
 | NanoHotpotQA | wordseg@th | 55.23 | 68.80 | 66.52 | 86.00 | 95.00 | 96.00 | 100 | 0 |
 | NanoMSMARCO | wordseg@th | 29.07 | 42.65 | 36.53 | 80.00 | 92.00 | 94.00 | 100-101 | 3 |
-| NanoNFCorpus | wordseg@th | 24.84 | 21.33 | 25.20 | 18.29 | 21.02 | 23.35 | 100-101 | 8 |
+| NanoNFCorpus | wordseg@th | 26.63 | 24.09 | 27.43 | 18.36 | 21.06 | 25.49 | 100-101 | 6 |
 | NanoNQ | wordseg@th | 31.91 | 53.67 | 42.46 | 84.00 | 92.00 | 93.00 | 100-101 | 3 |
 | NanoQuoraRetrieval | wordseg@th | 72.67 | 88.59 | 79.28 | 96.00 | 100.00 | 100.00 | 100 | 0 |
 | NanoSCIDOCS | wordseg@th | 26.41 | 29.15 | 31.65 | 55.97 | 60.07 | 62.47 | 100-101 | 2 |
@@ -555,7 +555,7 @@ Dense means `microsoft/harrier-oss-v1-270m` with the `web_search_query` prompt a
 
 ## Hybrid Safeguard Summary
 
-- Safeguard positives: 33
+- Safeguard positives: 31
 - Rows limited by corpus size: 0
 - Metadata file: `reranking_hybrid_metadata.json`
 

bm25/NanoNFCorpus-00000-of-00001.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7648253b06e212f27df1ae26bdc124e84a5201e4a735135e71b8780a935753b6
-size 301176
+oid sha256:70747b27545642649e73e5bfefd7d001991694ba413118e2ff768554de5f4f18
+size 301220

manifest.json

@@ -76,9 +76,9 @@
       "split_name": "NanoNFCorpus",
       "queries": 50,
       "corpus": 2953,
-      "qrels": 1651,
-      "forced_doc_count": 1397,
-      "bm25_ndcg_at_10": 0.32430076432568067,
+      "qrels": 2518,
+      "forced_doc_count": 1818,
+      "bm25_ndcg_at_10": 0.2663451597846342,
       "bm25_query_coverage": 1.0,
       "bm25_relevant_coverage": 1.0
     },

metadata/NanoNFCorpus.json

@@ -2,12 +2,14 @@
   "split_name": "NanoNFCorpus",
   "queries": 50,
   "corpus": 2953,
-  "qrels": 1651,
+  "qrels": 2518,
   "source_non_positive_qrels": 0,
   "qrels_selection": {
-    "qrels_per_query_cap": 100,
-    "removed_qrels_over_cap": 0,
-    "qrels_cap_policy": "positive qrels are capped per query to the BM25 candidate top_k"
+    "qrels_policy": "full positive qrels preserved",
+    "restored_qrels_from": "sentence-transformers/NanoBEIR-en@beb106fbcfaa599c508c667041bf8c85fd78736b",
+    "previous_hakari_qrels": 1651,
+    "restored_qrels": 867,
+    "max_positive_qrels_per_query": 463
   },
   "bm25": {
     "config": {
@@ -17,10 +19,10 @@
       "tokenizer": "wordseg",
       "tokenizer_name": "th",
       "stemmer_algorithm": "english",
-      "top_k": 100,
+      "top_k": 500,
       "k1": 1.5,
       "b": 0.75,
-      "show_progress": true,
+      "show_progress": false,
       "auto_selected": true,
       "auto_detected_language": "th",
       "auto_detection_language_counts": {
@@ -28,21 +30,55 @@
       },
       "auto_detection_sample_size": 10
     },
-    "ndcg_at_10": 0.32430076432568067,
+    "ndcg_at_10": 0.2663451597846342,
+    "recall_at_100": 0.18364903750285236,
     "candidate_coverage": {
-      "top_k": 100,
+      "top_k": 500,
       "query_count": 50,
       "query_with_relevance_count": 50,
       "covered_query_count": 50,
       "query_coverage": 1.0,
-      "relevant_count": 1651,
-      "covered_relevant_count": 1651,
+      "relevant_count": 2518,
+      "covered_relevant_count": 2518,
       "relevant_coverage": 1.0
     },
-    "rebuild_policy": "BM25 candidates were recomputed locally and qrels-positive documents missing from the raw top-k were forced into the final candidate list.",
-    "forced_queries": 48,
-    "forced_doc_count": 1397,
+    "rebuild_policy": "BM25 top-500 was recomputed locally; full qrels were preserved; qrels-positive documents were forced into the top-500 candidate list when missing.",
+    "forced_queries": 47,
+    "forced_doc_count": 1818,
     "missing_positive_doc_count_after_forcing": 0,
-    "rebuilt_at_utc": "2026-05-19T01:35:43.224847+00:00"
+    "rebuilt_at_utc": "2026-06-07T22:41:50.249844+00:00"
+  },
+  "harrier_oss_v1_270m": {
+    "source": "copied from current hakari-bench dataset; dense top-500 is qrels-independent",
+    "ndcg_at_10": 0.24094354685895156,
+    "recall_at_100": 0.21059042279243023,
+    "candidate_coverage": {
+      "top_k": 500,
+      "query_count": 50,
+      "query_with_relevance_count": 50,
+      "covered_query_count": 47,
+      "query_coverage": 0.94,
+      "relevant_count": 2518,
+      "covered_relevant_count": 782,
+      "relevant_coverage": 0.31056393963463064
+    }
+  },
+  "reranking_hybrid": {
+    "source": "rebuilt from BM25 top-500 and harrier_oss_v1_270m top-500 with RRF",
+    "rrf_k": 100,
+    "top_k": 100,
+    "ndcg_at_10": 0.27426706085714764,
+    "recall_at_100": 0.2549465786729181,
+    "candidate_coverage": {
+      "top_k": 100,
+      "query_count": 50,
+      "query_with_relevance_count": 50,
+      "covered_query_count": 44,
+      "query_coverage": 0.88,
+      "relevant_count": 2518,
+      "covered_relevant_count": 400,
+      "relevant_coverage": 0.15885623510722796
+    },
+    "safeguard_positive_count": 6
   }
 }

qrels/NanoNFCorpus-00000-of-00001.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:040edc79f4314589b75c857f37766b4f60e4cc3158728356617258096d549cd8
-size 9758
+oid sha256:40724452b59698b31db1a9de39b35400e99d18398a053a71d57ddf11d1d59a55
+size 13254

reranking_hybrid/NanoNFCorpus-00000-of-00001.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e96ef11ef29f9c964d935962009077fa1b0a3472ad86647ef250c23c6bfed515
-size 61929
+oid sha256:8489a3029dae1e609540bdbb8369778b310f09c81ebfd2a95de42dee3725d0aa
+size 61885