| layer_type | N | M | Q | alpha | D | alpha-hat | log_SN | % Rand | num_traps | num_fingers | rank_loss | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| layer_id | ||||||||||||
| 1 | EMBEDDING | 32128 | 512 | 62.75 | 2.28 | 0.07 | 18.77 | 8.22 | 74.06 | 0 | 4 | 0 |
| 8 | DENSE | 512 | 384 | 1.33 | 2.06 | 0.14 | 2.29 | 1.11 | 90.43 | 0 | 1 | 0 |
| 9 | DENSE | 512 | 384 | 1.33 | 1.97 | 0.13 | 5.74 | 2.92 | 91.16 | 0 | 1 | 0 |
| 10 | DENSE | 512 | 384 | 1.33 | 3.96 | 0.14 | 10.38 | 2.62 | 93.86 | 0 | 5 | 0 |
| 11 | DENSE | 512 | 384 | 1.33 | 3.15 | 0.03 | 10.98 | 3.48 | 84.97 | 0 | 0 | 0 |
| 17 | DENSE | 1024 | 512 | 2.00 | 3.19 | 0.05 | 10.65 | 3.34 | 88.06 | 0 | 0 | 0 |
| 18 | DENSE | 1024 | 512 | 2.00 | 3.58 | 0.08 | 11.66 | 3.26 | 91.92 | 0 | 0 | 0 |
| 19 | DENSE | 1024 | 512 | 2.00 | 3.08 | 0.03 | 11.14 | 3.61 | 82.80 | 0 | 0 | 0 |
| 28 | DENSE | 512 | 384 | 1.33 | 5.39 | 0.05 | 6.61 | 1.23 | 92.23 | 0 | 0 | 0 |
| 29 | DENSE | 512 | 384 | 1.33 | 5.20 | 0.05 | 15.46 | 2.97 | 92.62 | 0 | 0 | 0 |
| 30 | DENSE | 512 | 384 | 1.33 | 3.66 | 0.11 | 9.46 | 2.59 | 94.97 | 0 | 0 | 0 |
| 31 | DENSE | 512 | 384 | 1.33 | 3.30 | 0.04 | 12.29 | 3.73 | 81.74 | 0 | 0 | 0 |
| 36 | DENSE | 1024 | 512 | 2.00 | 3.53 | 0.03 | 11.74 | 3.33 | 89.18 | 0 | 0 | 0 |
| 37 | DENSE | 1024 | 512 | 2.00 | 4.92 | 0.08 | 16.02 | 3.26 | 94.95 | 0 | 0 | 0 |
| 38 | DENSE | 1024 | 512 | 2.00 | 3.19 | 0.02 | 11.26 | 3.53 | 84.84 | 0 | 0 | 0 |
| 47 | DENSE | 512 | 384 | 1.33 | 3.61 | 0.06 | 3.57 | 0.99 | 89.70 | 0 | 0 | 0 |
| 48 | DENSE | 512 | 384 | 1.33 | 5.01 | 0.05 | 14.44 | 2.88 | 91.78 | 0 | 0 | 0 |
| 49 | DENSE | 512 | 384 | 1.33 | 3.31 | 0.11 | 9.76 | 2.94 | 89.23 | 0 | 0 | 0 |
| 50 | DENSE | 512 | 384 | 1.33 | 4.03 | 0.04 | 14.45 | 3.58 | 83.18 | 0 | 0 | 0 |
| 55 | DENSE | 1024 | 512 | 2.00 | 3.34 | 0.04 | 11.13 | 3.33 | 87.25 | 0 | 0 | 0 |
| 56 | DENSE | 1024 | 512 | 2.00 | 6.37 | 0.06 | 22.15 | 3.48 | 94.34 | 0 | 0 | 0 |
| 57 | DENSE | 1024 | 512 | 2.00 | 3.53 | 0.04 | 13.16 | 3.72 | 86.28 | 0 | 0 | 0 |
| 66 | DENSE | 512 | 384 | 1.33 | 4.75 | 0.06 | 4.05 | 0.85 | 92.65 | 0 | 0 | 0 |
| 67 | DENSE | 512 | 384 | 1.33 | 5.44 | 0.06 | 14.78 | 2.72 | 93.45 | 0 | 0 | 0 |
| 68 | DENSE | 512 | 384 | 1.33 | 4.85 | 0.14 | 14.12 | 2.91 | 94.55 | 0 | 7 | 0 |
| 69 | DENSE | 512 | 384 | 1.33 | 4.09 | 0.05 | 13.71 | 3.35 | 91.18 | 0 | 0 | 0 |
| 74 | DENSE | 1024 | 512 | 2.00 | 3.41 | 0.04 | 11.20 | 3.29 | 86.15 | 0 | 0 | 0 |
| 75 | DENSE | 1024 | 512 | 2.00 | 5.67 | 0.10 | 20.22 | 3.57 | 93.66 | 0 | 1 | 0 |
| 76 | DENSE | 1024 | 512 | 2.00 | 3.69 | 0.04 | 13.47 | 3.65 | 88.10 | 0 | 0 | 0 |
| 85 | DENSE | 512 | 384 | 1.33 | 3.74 | 0.05 | 3.16 | 0.85 | 88.58 | 0 | 0 | 0 |
| 86 | DENSE | 512 | 384 | 1.33 | 3.44 | 0.10 | 9.19 | 2.67 | 88.25 | 0 | 1 | 0 |
| 87 | DENSE | 512 | 384 | 1.33 | 2.81 | 0.14 | 8.74 | 3.11 | 90.26 | 0 | 8 | 0 |
| 88 | DENSE | 512 | 384 | 1.33 | 4.20 | 0.05 | 16.10 | 3.83 | 85.70 | 0 | 0 | 0 |
| 93 | DENSE | 1024 | 512 | 2.00 | 3.48 | 0.06 | 11.22 | 3.23 | 85.54 | 0 | 0 | 0 |
| 94 | DENSE | 1024 | 512 | 2.00 | 5.16 | 0.07 | 19.03 | 3.69 | 92.96 | 0 | 0 | 0 |
| 95 | DENSE | 1024 | 512 | 2.00 | 3.88 | 0.04 | 14.62 | 3.76 | 87.85 | 0 | 0 | 0 |
| 104 | DENSE | 512 | 384 | 1.33 | 4.77 | 0.04 | 4.22 | 0.88 | 87.65 | 0 | 0 | 0 |
| 105 | DENSE | 512 | 384 | 1.33 | 2.16 | 0.14 | 5.67 | 2.62 | 90.76 | 0 | 8 | 0 |
| 106 | DENSE | 512 | 384 | 1.33 | 4.63 | 0.11 | 14.27 | 3.08 | 92.25 | 0 | 0 | 0 |
| 107 | DENSE | 512 | 384 | 1.33 | 4.00 | 0.06 | 15.14 | 3.79 | 86.05 | 0 | 0 | 0 |
| 112 | DENSE | 1024 | 512 | 2.00 | 3.24 | 0.10 | 10.25 | 3.16 | 85.37 | 0 | 6 | 0 |
| 113 | DENSE | 1024 | 512 | 2.00 | 5.38 | 0.07 | 20.53 | 3.82 | 91.77 | 0 | 0 | 0 |
| 114 | DENSE | 1024 | 512 | 2.00 | 3.91 | 0.04 | 14.64 | 3.74 | 88.24 | 0 | 0 | 0 |
| 123 | DENSE | 512 | 384 | 1.33 | 3.73 | 0.05 | 3.81 | 1.02 | 85.99 | 0 | 0 | 0 |
| 124 | DENSE | 512 | 384 | 1.33 | 3.50 | 0.07 | 9.39 | 2.68 | 88.38 | 0 | 0 | 0 |
| 125 | DENSE | 512 | 384 | 1.33 | 4.08 | 0.13 | 13.32 | 3.26 | 90.21 | 0 | 3 | 0 |
| 126 | DENSE | 512 | 384 | 1.33 | 3.09 | 0.10 | 12.08 | 3.91 | 84.91 | 0 | 1 | 0 |
| 131 | DENSE | 1024 | 512 | 2.00 | 2.87 | 0.08 | 8.87 | 3.08 | 85.35 | 0 | 1 | 0 |
| 132 | DENSE | 1024 | 512 | 2.00 | 4.54 | 0.05 | 17.83 | 3.93 | 91.15 | 0 | 0 | 0 |
| 133 | DENSE | 1024 | 512 | 2.00 | 3.98 | 0.03 | 14.52 | 3.65 | 89.08 | 0 | 0 | 0 |
| 142 | DENSE | 512 | 384 | 1.33 | 2.91 | 0.04 | 2.47 | 0.85 | 80.59 | 0 | 0 | 0 |
| 143 | DENSE | 512 | 384 | 1.33 | 2.57 | 0.07 | 6.36 | 2.48 | 83.88 | 0 | 0 | 0 |
| 144 | DENSE | 512 | 384 | 1.33 | 4.82 | 0.12 | 15.94 | 3.30 | 91.48 | 0 | 1 | 0 |
| 145 | DENSE | 512 | 384 | 1.33 | 4.25 | 0.06 | 17.23 | 4.06 | 83.27 | 0 | 0 | 0 |
| 150 | DENSE | 1024 | 512 | 2.00 | 2.80 | 0.09 | 8.71 | 3.12 | 85.16 | 0 | 2 | 0 |
| 151 | DENSE | 1024 | 512 | 2.00 | 4.39 | 0.05 | 18.08 | 4.12 | 91.25 | 1 | 0 | 0 |
| 152 | DENSE | 1024 | 512 | 2.00 | 3.46 | 0.04 | 13.14 | 3.80 | 86.15 | 0 | 0 | 0 |
| 165 | DENSE | 512 | 384 | 1.33 | 3.90 | 0.04 | 5.10 | 1.31 | 86.20 | 0 | 0 | 0 |
| 166 | DENSE | 512 | 384 | 1.33 | 4.57 | 0.08 | 13.71 | 3.00 | 88.79 | 0 | 0 | 0 |
| 167 | DENSE | 512 | 384 | 1.33 | 4.97 | 0.07 | 13.52 | 2.72 | 90.84 | 0 | 0 | 0 |
| 168 | DENSE | 512 | 384 | 1.33 | 3.20 | 0.06 | 10.59 | 3.31 | 83.44 | 0 | 0 | 0 |
| 174 | DENSE | 512 | 384 | 1.33 | 4.21 | 0.04 | 5.12 | 1.22 | 91.51 | 0 | 0 | 0 |
| 175 | DENSE | 512 | 384 | 1.33 | 3.36 | 0.09 | 10.10 | 3.01 | 84.17 | 0 | 0 | 0 |
| 176 | DENSE | 512 | 384 | 1.33 | 5.18 | 0.04 | 13.19 | 2.55 | 91.97 | 0 | 0 | 0 |
| 177 | DENSE | 512 | 384 | 1.33 | 3.02 | 0.04 | 10.58 | 3.51 | 79.19 | 0 | 0 | 0 |
| 182 | DENSE | 1024 | 512 | 2.00 | 2.82 | 0.03 | 10.69 | 3.79 | 82.85 | 0 | 0 | 0 |
| 183 | DENSE | 1024 | 512 | 2.00 | 3.56 | 0.04 | 11.43 | 3.21 | 91.56 | 0 | 0 | 0 |
| 184 | DENSE | 1024 | 512 | 2.00 | 3.28 | 0.03 | 12.37 | 3.77 | 82.60 | 0 | 0 | 0 |
| 193 | DENSE | 512 | 384 | 1.33 | 2.91 | 0.04 | 4.13 | 1.42 | 82.05 | 0 | 0 | 0 |
| 194 | DENSE | 512 | 384 | 1.33 | 3.11 | 0.04 | 9.10 | 2.93 | 83.17 | 0 | 0 | 0 |
| 195 | DENSE | 512 | 384 | 1.33 | 3.63 | 0.05 | 11.04 | 3.04 | 87.86 | 0 | 0 | 0 |
| 196 | DENSE | 512 | 384 | 1.33 | 2.84 | 0.02 | 11.19 | 3.94 | 74.90 | 0 | 0 | 0 |
| 201 | DENSE | 512 | 384 | 1.33 | 4.11 | 0.05 | 5.22 | 1.27 | 88.53 | 0 | 0 | 0 |
| 202 | DENSE | 512 | 384 | 1.33 | 3.62 | 0.04 | 10.71 | 2.95 | 89.60 | 0 | 0 | 0 |
| 203 | DENSE | 512 | 384 | 1.33 | 3.54 | 0.08 | 9.64 | 2.72 | 89.91 | 0 | 0 | 0 |
| 204 | DENSE | 512 | 384 | 1.33 | 2.95 | 0.03 | 11.15 | 3.79 | 75.44 | 0 | 0 | 0 |
| 209 | DENSE | 1024 | 512 | 2.00 | 3.90 | 0.03 | 13.05 | 3.35 | 88.94 | 0 | 0 | 0 |
| 210 | DENSE | 1024 | 512 | 2.00 | 4.48 | 0.07 | 15.22 | 3.40 | 92.78 | 0 | 0 | 0 |
| 211 | DENSE | 1024 | 512 | 2.00 | 3.76 | 0.05 | 15.29 | 4.06 | 83.09 | 0 | 0 | 0 |
| 220 | DENSE | 512 | 384 | 1.33 | 3.09 | 0.05 | 2.93 | 0.95 | 83.03 | 0 | 0 | 0 |
| 221 | DENSE | 512 | 384 | 1.33 | 3.43 | 0.05 | 9.39 | 2.74 | 85.40 | 0 | 0 | 0 |
| 222 | DENSE | 512 | 384 | 1.33 | 3.86 | 0.12 | 11.74 | 3.04 | 88.54 | 0 | 1 | 0 |
| 223 | DENSE | 512 | 384 | 1.33 | 2.85 | 0.04 | 11.88 | 4.17 | 73.12 | 0 | 0 | 0 |
| 228 | DENSE | 512 | 384 | 1.33 | 4.70 | 0.06 | 5.59 | 1.19 | 91.37 | 0 | 0 | 0 |
| 229 | DENSE | 512 | 384 | 1.33 | 3.98 | 0.05 | 11.44 | 2.88 | 90.52 | 0 | 0 | 0 |
| 230 | DENSE | 512 | 384 | 1.33 | 4.39 | 0.11 | 12.70 | 2.89 | 94.06 | 0 | 1 | 0 |
| 231 | DENSE | 512 | 384 | 1.33 | 3.11 | 0.05 | 12.43 | 4.00 | 77.90 | 0 | 0 | 0 |
| 236 | DENSE | 1024 | 512 | 2.00 | 3.63 | 0.05 | 11.90 | 3.28 | 87.26 | 0 | 0 | 0 |
| 237 | DENSE | 1024 | 512 | 2.00 | 4.67 | 0.04 | 17.81 | 3.81 | 91.92 | 0 | 0 | 0 |
| 238 | DENSE | 1024 | 512 | 2.00 | 3.82 | 0.05 | 17.33 | 4.53 | 78.68 | 1 | 0 | 0 |
| 247 | DENSE | 512 | 384 | 1.33 | 3.78 | 0.07 | 3.54 | 0.94 | 86.32 | 0 | 0 | 0 |
| 248 | DENSE | 512 | 384 | 1.33 | 3.36 | 0.05 | 9.07 | 2.70 | 86.97 | 0 | 0 | 0 |
| 249 | DENSE | 512 | 384 | 1.33 | 6.01 | 0.10 | 19.50 | 3.24 | 91.86 | 0 | 0 | 0 |
| 250 | DENSE | 512 | 384 | 1.33 | 3.15 | 0.09 | 13.93 | 4.43 | 74.34 | 0 | 0 | 0 |
| 255 | DENSE | 512 | 384 | 1.33 | 2.78 | 0.05 | 3.78 | 1.36 | 84.01 | 0 | 0 | 0 |
| 256 | DENSE | 512 | 384 | 1.33 | 2.73 | 0.05 | 8.96 | 3.28 | 79.52 | 0 | 0 | 0 |
| 257 | DENSE | 512 | 384 | 1.33 | 7.44 | 0.05 | 22.28 | 2.99 | 94.89 | 0 | 0 | 0 |
| 258 | DENSE | 512 | 384 | 1.33 | 3.34 | 0.06 | 13.60 | 4.07 | 75.64 | 0 | 0 | 0 |
| 263 | DENSE | 1024 | 512 | 2.00 | 2.92 | 0.05 | 9.31 | 3.19 | 85.06 | 0 | 0 | 0 |
| 264 | DENSE | 1024 | 512 | 2.00 | 4.09 | 0.09 | 15.56 | 3.81 | 91.44 | 0 | 1 | 0 |
| 265 | DENSE | 1024 | 512 | 2.00 | 3.65 | 0.02 | 16.55 | 4.54 | 80.08 | 0 | 0 | 0 |
| 274 | DENSE | 512 | 384 | 1.33 | 4.98 | 0.04 | 5.69 | 1.14 | 89.00 | 0 | 0 | 0 |
| 275 | DENSE | 512 | 384 | 1.33 | 3.65 | 0.14 | 9.70 | 2.66 | 92.18 | 0 | 8 | 0 |
| 276 | DENSE | 512 | 384 | 1.33 | 5.15 | 0.12 | 16.62 | 3.23 | 91.85 | 0 | 1 | 0 |
| 277 | DENSE | 512 | 384 | 1.33 | 3.05 | 0.04 | 14.09 | 4.63 | 73.89 | 0 | 0 | 0 |
| 282 | DENSE | 512 | 384 | 1.33 | 3.30 | 0.05 | 4.55 | 1.38 | 84.37 | 0 | 0 | 0 |
| 283 | DENSE | 512 | 384 | 1.33 | 3.06 | 0.04 | 10.19 | 3.33 | 83.29 | 0 | 0 | 0 |
| 284 | DENSE | 512 | 384 | 1.33 | 3.69 | 0.14 | 11.61 | 3.15 | 91.92 | 0 | 5 | 0 |
| 285 | DENSE | 512 | 384 | 1.33 | 3.93 | 0.06 | 15.80 | 4.02 | 82.21 | 0 | 0 | 0 |
| 290 | DENSE | 1024 | 512 | 2.00 | 3.31 | 0.05 | 10.04 | 3.03 | 86.29 | 0 | 0 | 0 |
| 291 | DENSE | 1024 | 512 | 2.00 | 3.58 | 0.11 | 13.75 | 3.84 | 92.23 | 0 | 4 | 0 |
| 292 | DENSE | 1024 | 512 | 2.00 | 3.60 | 0.04 | 16.21 | 4.50 | 80.80 | 0 | 0 | 0 |
| 301 | DENSE | 512 | 384 | 1.33 | 3.67 | 0.04 | 4.86 | 1.33 | 85.16 | 0 | 0 | 0 |
| 302 | DENSE | 512 | 384 | 1.33 | 5.73 | 0.07 | 14.51 | 2.53 | 91.01 | 0 | 0 | 0 |
| 303 | DENSE | 512 | 384 | 1.33 | 4.88 | 0.13 | 16.57 | 3.39 | 94.92 | 0 | 5 | 0 |
| 304 | DENSE | 512 | 384 | 1.33 | 1.91 | 0.08 | 8.57 | 4.48 | 78.29 | 0 | 1 | 0 |
| 309 | DENSE | 512 | 384 | 1.33 | 4.24 | 0.06 | 5.06 | 1.19 | 88.65 | 0 | 0 | 0 |
| 310 | DENSE | 512 | 384 | 1.33 | 4.75 | 0.06 | 14.04 | 2.96 | 88.77 | 0 | 0 | 0 |
| 311 | DENSE | 512 | 384 | 1.33 | 2.60 | 0.11 | 8.58 | 3.30 | 92.70 | 0 | 0 | 0 |
| 312 | DENSE | 512 | 384 | 1.33 | 3.30 | 0.08 | 14.74 | 4.46 | 77.28 | 0 | 0 | 0 |
| 317 | DENSE | 1024 | 512 | 2.00 | 3.14 | 0.05 | 9.01 | 2.87 | 86.04 | 0 | 0 | 0 |
| 318 | DENSE | 1024 | 512 | 2.00 | 4.55 | 0.09 | 18.00 | 3.96 | 92.32 | 0 | 1 | 0 |
| 319 | DENSE | 1024 | 512 | 2.00 | 3.94 | 0.04 | 17.16 | 4.36 | 83.85 | 0 | 0 | 0 |
| 328 | DENSE | 512 | 384 | 1.33 | 4.06 | 0.07 | 2.81 | 0.69 | 89.98 | 0 | 0 | 0 |
| 329 | DENSE | 512 | 384 | 1.33 | 4.10 | 0.07 | 10.14 | 2.47 | 89.10 | 0 | 0 | 0 |
| 330 | DENSE | 512 | 384 | 1.33 | 2.36 | 0.14 | 7.90 | 3.35 | 94.70 | 0 | 2 | 0 |
| 331 | DENSE | 512 | 384 | 1.33 | 4.09 | 0.10 | 20.10 | 4.91 | 75.87 | 0 | 0 | 0 |
| 336 | DENSE | 512 | 384 | 1.33 | 3.62 | 0.06 | 4.43 | 1.22 | 83.47 | 0 | 0 | 0 |
| 337 | DENSE | 512 | 384 | 1.33 | 2.04 | 0.11 | 6.21 | 3.05 | 83.93 | 0 | 4 | 0 |
| 338 | DENSE | 512 | 384 | 1.33 | 10.00 | 0.11 | 33.78 | 3.38 | 95.50 | 0 | 0 | 0 |
| 339 | DENSE | 512 | 384 | 1.33 | 3.83 | 0.11 | 17.76 | 4.63 | 81.64 | 0 | 0 | 0 |
| 344 | DENSE | 1024 | 512 | 2.00 | 3.38 | 0.05 | 9.37 | 2.78 | 85.78 | 0 | 0 | 0 |
| 345 | DENSE | 1024 | 512 | 2.00 | 4.20 | 0.07 | 17.81 | 4.24 | 90.86 | 0 | 0 | 0 |
| 346 | DENSE | 1024 | 512 | 2.00 | 3.92 | 0.05 | 17.31 | 4.42 | 85.35 | 0 | 0 | 0 |
| 355 | DENSE | 512 | 384 | 1.33 | 4.61 | 0.03 | 3.94 | 0.86 | 92.82 | 0 | 0 | 0 |
| 356 | DENSE | 512 | 384 | 1.33 | 4.69 | 0.05 | 13.03 | 2.78 | 92.94 | 0 | 0 | 0 |
| 357 | DENSE | 512 | 384 | 1.33 | 7.66 | 0.06 | 27.15 | 3.55 | 93.97 | 0 | 0 | 0 |
| 358 | DENSE | 512 | 384 | 1.33 | 3.26 | 0.07 | 14.32 | 4.39 | 77.52 | 0 | 0 | 0 |
| 363 | DENSE | 512 | 384 | 1.33 | 2.83 | 0.09 | 4.41 | 1.56 | 75.52 | 0 | 0 | 0 |
| 364 | DENSE | 512 | 384 | 1.33 | 1.94 | 0.09 | 6.44 | 3.32 | 77.55 | 0 | 1 | 0 |
| 365 | DENSE | 512 | 384 | 1.33 | 6.62 | 0.13 | 25.27 | 3.82 | 88.86 | 0 | 1 | 0 |
| 366 | DENSE | 512 | 384 | 1.33 | 3.68 | 0.06 | 17.22 | 4.68 | 85.99 | 0 | 0 | 0 |
| 371 | DENSE | 1024 | 512 | 2.00 | 3.24 | 0.04 | 10.59 | 3.27 | 81.62 | 0 | 0 | 0 |
| 372 | DENSE | 1024 | 512 | 2.00 | 3.10 | 0.08 | 15.34 | 4.95 | 85.96 | 0 | 0 | 0 |
| 373 | DENSE | 1024 | 512 | 2.00 | 2.86 | 0.05 | 15.16 | 5.30 | 71.02 | 5 | 0 | 0 |





































































































































































































































































































































































































































































































































































































































































































































































































































































































