TNLG v2 (6.7B): benchmark results
Provider: Other.
Aggregate Rating 1111 ± 33, rank #10604 of 10916 rated models, from 32 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| HELM Classic - BLiMP | 83.47 | Exact Match (%) | 93.5 |
| HELM Classic - Entity Data Imputation | 82.62 | Exact Match (%) | 77.3 |
| HELM Classic - Synthetic Reasoning Natural | 24.69 | F1 (%) | 75 |
| HELM Classic - CNN/DailyMail | 14.62 | ROUGE-2 (%) | 68.3 |
| HELM Classic - Synthetic Reasoning Abstract | 23.24 | Exact Match (%) | 58.8 |
| HELM Classic - Dyck | 55.4 | Exact Match (%) | 52.9 |
| HELM Classic - XSUM | 11.01 | ROUGE-2 (%) | 48.8 |
| HELM Classic - QuAC | 34.48 | F1 (%) | 46.2 |
| HELM Classic - WikiFact | 23.61 | Exact Match (%) | 45.5 |
| HELM Classic - NarrativeQA | 63.09 | F1 (%) | 43.1 |
| HELM Classic - CivilComments | 53.15 | Exact Match (%) | 37.9 |
| HELM Classic - IMDB | 92.73 | Exact Match (%) | 37.9 |
Interactive version: theaggregate.ai/model?slug=tnlg-v2-6-7b · How It Works · Data refreshed daily, snapshot 2026-11-08.