--- license: llama3.1 --- | | Wiki | C4 | PIQA | ARC-E | ARC-C | HellaSwag | Wino | Avg. | | ----------- | ---- | ----- | ----- | ----- | ----- | --------- | ----- | ----- | | Unquantized | 6.31 | 9.60 | 79.16 | 82.20 | 52.65 | 60.71 | 74.03 | 69.75 | | W4G64 | 6.24 | 10.06 | 79.38 | 81.61 | 51.54 | 59.57 | 73.56 | 69.13 | | W3G64 | 7.23 | 11.83 | 77.91 | 76.98 | 46.33 | 56.74 | 70.32 | 65.66 | Revisions available in this repository: - `main` (W4G64, scales learned); - `nfl_w3g64` (W3G64, scales learned); Evaluations are provided for models with learned scales.
Benchmark scores (zero-shot) are computed with [`lm-evaluation-harness`](https://github.com/EleutherAI/lm-evaluation-harness).