Spaces:
Running
Running
| 5-shot,FineWeb-1.5T,Ours-Base,Ours-Upsampling2,All-Upsampling1 | |
| time: 76 min,Llama-8x8B-seq8192,Llama-8x8B-seq8192,Llama-8x8B-seq8192,Llama-8x8B-seq8192 | |
| 5k,0.1025,0.1232,0.126, | |
| 10k,0.2073,,0.115,0.2604 | |
| 15k,0.3005,,0.1872,0.3244 | |
| 20k,0.3506,0.2795,0.2719,0.3637 | |
| 25k,0.307,,0.4093,0.412 | |
| 30k,0.2461,0.2974,0.4195,0.4294 | |
| 35k,0.3639,0.3572,0.3587,0.4428 | |
| 40k,0.3537,0.0346,0.4434,0.4623 | |
| 45k,0.3602,0.2674,0.4366,0.4792 | |
| 50k,0.2407,0.3689,0.4051,0.4795 | |
| 55k,0.2081,0.4101,0.323,0.494 | |
| 60k,0.4068,0.4107,0.4469,0.513 | |
| 65k,0.3145,0.4477,0.4907,0.5087 | |
| 70k,0.4102,0.4736,0.492,0.5129 | |
| 75k,0.282,0.4226,0.2245,0.5042 | |
| 80k,0.0975,0.4217,,0.5301 | |
| 85k,0.0722,0.4763,0.5029,0.535 | |
| 90k,0.3388,0.1472,0.0317,0.522 | |
| 95k,0.5283,0.4938,0.518,0.5446 | |
| 100k,0.4317,0.11,0.5358,0.5514 | |
| 105k,0.1886,,0.5153,0.5562 | |
| 110k,0.351,,0.5182,0.5654 | |
| 115k,0.3692,0.4759,0.5132,0.5577 | |
| 120k,0.369,0.4352,0.5483,0.5658 | |
| 125k,0.3365,0.5206,0.5211,0.5658 | |
| 130k,0.355,0.0088,0.5245,0.5609 | |
| 135k,0.3892,,0.3977,0.5774 | |
| 140k,0.393,,0.4991,0.5675 | |
| 145k,0.4538,,0.4872,0.5639 | |
| 150k,0.2883,,,0.5844 | |
| 155k,0.4185,,0.1586,0.5755 | |
| 160k,0.272,,0.563,0.5864 | |
| 165k,0.4252,,0.5642,0.5853 | |
| 170k,0.1507,,0.5739, | |
| 175k,0.3242,,0.564,0.5979 | |
| 180k,0.2653,,0.5912,0.6054 | |
| 185k,0.2651,,0.5852,0.6064 | |
| 190k,0.238,,,0.5996 | |
| 195k,0.4048,,,0.6243 | |
| 200k,0.5058,,,0.6248 | |
| 205k,0.0945,,,0.6224 | |
| 210k,0.1557,,,0.6311 | |
| 215k,0.2483,,,0.6293 | |
| 220k,0.1725,,,0.6375 | |
| 225k,0.2467,,,0.634 | |
| 230k,0.1653,,,0.6436 | |
| 235k,0.1884,,,0.6411 | |
| 240k,0.0719,,, | |
| 245k,0.3757,,, | |
| 250k,0.5859,,, | |
| 255k,0.4987,,, | |
| 260k,0.394,,, | |
| 265k,0.3607,,, | |
| 270k,0.3898,,, | |
| 275k,0.4123,,, | |
| 280k,0.2413,,, | |
| 285k,0.3665,,, | |
| 290k,,,, | |
| 300k,,,, | |
| 305k,,,, | |
| 310k,,,, | |
| 315k,,,, | |
| 320k,,,, | |
| 325k,,,, | |
| 330k,,,, | |
| 335k,,,, |