The dataset viewer is not available because its heuristics could not detect any supported data files. You can try uploading some data files, or configuring the data files location manually.
ViNID results
Verified name/diacritic evaluation artifacts. Full, pilot, context extension, repeats and controls are separate.
Protocol, current tables and findings.
Numeric source revision: 445b039848c4436b81ff24baac3b0d5e8a4a2119. Use fixed commits when retrieving scores.
Folder rules
benchmark-v3/
index.json
datasets/main/ and datasets/context-extension/
backbones/<backbone>/E1|E2|E3/main|context-extension/
full/pass-01|pass-02|pass-03/
pilot/calibration-01/
metrics.json
predictions.jsonl
requests.jsonl
provenance/
controls/<control>/E1|E2|E3/main/full/pass-01/
analysis/repeats/
analysis/context-comparison/
analysis/paired-e1-e2/
analysis/metric-sensitivity/
experiments.md
Only verified leaves exist. Run IDs remain metadata. The index stores hashes and original-file mappings. Google/Bing identify services; their internal checkpoints are undisclosed. Tier codes 85/25 are benchmark strata, not population percentages. E1/E2 measure choices under ambiguous ASCII input; E3 measures supplied-name preservation. Source scoring/data are unchanged. Checks, errors and incomplete runs are not model scores.
Historical legacy runs/indexes remain separate, at original revision
abd31d35632ee85df843c46cf7bc87ae4b139b88; they are not benchmark-v3 scores.
- Downloads last month
- 458