{
  "checked_at": "2026-10-09",
  "source": "https://huggingface.co/SlayerLab/gollem-v5-ckpts/blob/963440ce6ab4ada7e95da4c1faa28ebb00c082d4/README.md",
  "source_revision": "963440ce6ab4ada7e95da4c1faa28ebb00c082d4",
  "source_sha256": "6f70a32c669f73038790194b9177321e13daa5501f22aac00123540d611b8c8f",
  "local_input": "scaling_laws/data/models.csv",
  "provenance": "Size comparison from locally recorded model results, checked against the public model card; checkpoint series transcribed from the same pinned card. Reported measurements, not a new benchmark run.",
  "protocol": "Glint-1.3 benchmark.py: BLiMP 67,000 pairs; ARC-Easy 2,376 questions; WikiText-2 test. One training seed.",
  "interpretation": "Training curves at fixed model size; cross-size endpoints also differ in architecture, recipe and token budgets. No fitted power law or extrapolation.",
  "models": [
    {
      "model": "GoLLeM-v5 16M",
      "size_label": "16M",
      "params_m": 17.4,
      "training_tokens_b": 16.0,
      "blimp_pct": 70.08,
      "arc_easy_pct": 40.91,
      "wikitext2_byte_ppl": 2.6746,
      "url": "https://huggingface.co/SlayerLab/gollem-v5-ckpts/tree/963440ce6ab4ada7e95da4c1faa28ebb00c082d4/run_16m_expanded"
    },
    {
      "model": "GoLLeM-v5 32M",
      "size_label": "32M",
      "params_m": 31.6,
      "training_tokens_b": 16.0,
      "blimp_pct": 73.48,
      "arc_easy_pct": 44.44,
      "wikitext2_byte_ppl": 2.5386,
      "url": "https://huggingface.co/SlayerLab/gollem-v5-ckpts/tree/963440ce6ab4ada7e95da4c1faa28ebb00c082d4/run_32m_16b"
    },
    {
      "model": "GoLLeM-v5 64M",
      "size_label": "64M",
      "params_m": 62.9,
      "training_tokens_b": 24.9,
      "blimp_pct": 76.16,
      "arc_easy_pct": 48.19,
      "wikitext2_byte_ppl": 2.3474,
      "url": "https://huggingface.co/SlayerLab/gollem-v5-ckpts/tree/963440ce6ab4ada7e95da4c1faa28ebb00c082d4/final_64m_14x576"
    },
    {
      "model": "GoLLeM-v5 128M",
      "size_label": "128M",
      "params_m": 122.8,
      "training_tokens_b": 24.9,
      "blimp_pct": 79.09,
      "arc_easy_pct": 53.24,
      "wikitext2_byte_ppl": 2.2538,
      "url": "https://huggingface.co/SlayerLab/gollem-v5-ckpts/tree/963440ce6ab4ada7e95da4c1faa28ebb00c082d4/final_128m_16x768"
    }
  ],
  "checkpoints": [
    {
      "model": "64M",
      "step": 20000,
      "training_tokens": 655360000,
      "arc_easy_pct": 40.61,
      "blimp_pct": 72.83,
      "wikitext2_byte_ppl": 2.607,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 40000,
      "training_tokens": 1310720000,
      "arc_easy_pct": 43.43,
      "blimp_pct": 73.46,
      "wikitext2_byte_ppl": 2.5388,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 60000,
      "training_tokens": 1966080000,
      "arc_easy_pct": 43.77,
      "blimp_pct": 73.41,
      "wikitext2_byte_ppl": 2.4992,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 80000,
      "training_tokens": 2621440000,
      "arc_easy_pct": 45.08,
      "blimp_pct": 73.52,
      "wikitext2_byte_ppl": 2.4777,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 100000,
      "training_tokens": 3276800000,
      "arc_easy_pct": 45.5,
      "blimp_pct": 74.49,
      "wikitext2_byte_ppl": 2.4733,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 120000,
      "training_tokens": 3932160000,
      "arc_easy_pct": 45.58,
      "blimp_pct": 75.15,
      "wikitext2_byte_ppl": 2.4593,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 140000,
      "training_tokens": 4587520000,
      "arc_easy_pct": 44.7,
      "blimp_pct": 74.74,
      "wikitext2_byte_ppl": 2.4501,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 160000,
      "training_tokens": 5242880000,
      "arc_easy_pct": 45.79,
      "blimp_pct": 73.81,
      "wikitext2_byte_ppl": 2.4416,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 180000,
      "training_tokens": 5898240000,
      "arc_easy_pct": 44.99,
      "blimp_pct": 73.72,
      "wikitext2_byte_ppl": 2.4344,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 200000,
      "training_tokens": 6553600000,
      "arc_easy_pct": 45.5,
      "blimp_pct": 74.72,
      "wikitext2_byte_ppl": 2.4302,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 220000,
      "training_tokens": 7208960000,
      "arc_easy_pct": 46.3,
      "blimp_pct": 74.38,
      "wikitext2_byte_ppl": 2.4184,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 240000,
      "training_tokens": 7864320000,
      "arc_easy_pct": 46.46,
      "blimp_pct": 74.42,
      "wikitext2_byte_ppl": 2.4185,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 260000,
      "training_tokens": 8519680000,
      "arc_easy_pct": 46.3,
      "blimp_pct": 74.26,
      "wikitext2_byte_ppl": 2.4077,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 280000,
      "training_tokens": 9175040000,
      "arc_easy_pct": 46.51,
      "blimp_pct": 74.48,
      "wikitext2_byte_ppl": 2.4016,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 300000,
      "training_tokens": 9830400000,
      "arc_easy_pct": 46.59,
      "blimp_pct": 74.41,
      "wikitext2_byte_ppl": 2.4003,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 320000,
      "training_tokens": 10485760000,
      "arc_easy_pct": 46.97,
      "blimp_pct": 74.56,
      "wikitext2_byte_ppl": 2.3955,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 340000,
      "training_tokens": 11141120000,
      "arc_easy_pct": 46.8,
      "blimp_pct": 75.05,
      "wikitext2_byte_ppl": 2.3987,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 360000,
      "training_tokens": 11796480000,
      "arc_easy_pct": 45.79,
      "blimp_pct": 75.26,
      "wikitext2_byte_ppl": 2.3917,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 380000,
      "training_tokens": 12451840000,
      "arc_easy_pct": 46.72,
      "blimp_pct": 75.14,
      "wikitext2_byte_ppl": 2.3957,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 400000,
      "training_tokens": 13107200000,
      "arc_easy_pct": 46.97,
      "blimp_pct": 75.31,
      "wikitext2_byte_ppl": 2.393,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 420000,
      "training_tokens": 13762560000,
      "arc_easy_pct": 47.56,
      "blimp_pct": 75.46,
      "wikitext2_byte_ppl": 2.3865,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 440000,
      "training_tokens": 14417920000,
      "arc_easy_pct": 46.17,
      "blimp_pct": 75.74,
      "wikitext2_byte_ppl": 2.378,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 460000,
      "training_tokens": 15073280000,
      "arc_easy_pct": 46.3,
      "blimp_pct": 75.47,
      "wikitext2_byte_ppl": 2.3759,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 480000,
      "training_tokens": 15728640000,
      "arc_easy_pct": 46.72,
      "blimp_pct": 75.69,
      "wikitext2_byte_ppl": 2.3734,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 500000,
      "training_tokens": 16384000000,
      "arc_easy_pct": 47.05,
      "blimp_pct": 75.55,
      "wikitext2_byte_ppl": 2.3693,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 520000,
      "training_tokens": 17039360000,
      "arc_easy_pct": 46.8,
      "blimp_pct": 75.6,
      "wikitext2_byte_ppl": 2.3674,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 540000,
      "training_tokens": 17694720000,
      "arc_easy_pct": 46.46,
      "blimp_pct": 75.83,
      "wikitext2_byte_ppl": 2.3678,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 560000,
      "training_tokens": 18350080000,
      "arc_easy_pct": 47.81,
      "blimp_pct": 76.17,
      "wikitext2_byte_ppl": 2.3648,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 580000,
      "training_tokens": 19005440000,
      "arc_easy_pct": 47.47,
      "blimp_pct": 76.38,
      "wikitext2_byte_ppl": 2.3604,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 600000,
      "training_tokens": 19660800000,
      "arc_easy_pct": 47.56,
      "blimp_pct": 76.11,
      "wikitext2_byte_ppl": 2.3561,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 620000,
      "training_tokens": 20316160000,
      "arc_easy_pct": 48.78,
      "blimp_pct": 75.96,
      "wikitext2_byte_ppl": 2.3589,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 640000,
      "training_tokens": 20971520000,
      "arc_easy_pct": 48.15,
      "blimp_pct": 76.64,
      "wikitext2_byte_ppl": 2.3538,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 660000,
      "training_tokens": 21626880000,
      "arc_easy_pct": 47.39,
      "blimp_pct": 76.69,
      "wikitext2_byte_ppl": 2.3557,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 680000,
      "training_tokens": 22282240000,
      "arc_easy_pct": 48.02,
      "blimp_pct": 75.96,
      "wikitext2_byte_ppl": 2.3529,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 700000,
      "training_tokens": 22937600000,
      "arc_easy_pct": 48.19,
      "blimp_pct": 76.07,
      "wikitext2_byte_ppl": 2.3513,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 720000,
      "training_tokens": 23592960000,
      "arc_easy_pct": 48.61,
      "blimp_pct": 75.93,
      "wikitext2_byte_ppl": 2.3489,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 740000,
      "training_tokens": 24248320000,
      "arc_easy_pct": 47.98,
      "blimp_pct": 76.16,
      "wikitext2_byte_ppl": 2.3475,
      "status": "intermediate"
    },
    {
      "model": "64M",
      "step": 760000,
      "training_tokens": 24903680000,
      "arc_easy_pct": 48.19,
      "blimp_pct": 76.16,
      "wikitext2_byte_ppl": 2.3474,
      "status": "final"
    },
    {
      "model": "128M",
      "step": 20000,
      "training_tokens": 655360000,
      "arc_easy_pct": 43.9,
      "blimp_pct": 73.91,
      "wikitext2_byte_ppl": 2.5383,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 40000,
      "training_tokens": 1310720000,
      "arc_easy_pct": 46.68,
      "blimp_pct": 74.81,
      "wikitext2_byte_ppl": 2.4612,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 60000,
      "training_tokens": 1966080000,
      "arc_easy_pct": 47.43,
      "blimp_pct": 75.24,
      "wikitext2_byte_ppl": 2.4151,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 80000,
      "training_tokens": 2621440000,
      "arc_easy_pct": 46.59,
      "blimp_pct": 77.57,
      "wikitext2_byte_ppl": 2.3959,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 100000,
      "training_tokens": 3276800000,
      "arc_easy_pct": 47.81,
      "blimp_pct": 77.35,
      "wikitext2_byte_ppl": 2.3813,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 120000,
      "training_tokens": 3932160000,
      "arc_easy_pct": 48.82,
      "blimp_pct": 77.51,
      "wikitext2_byte_ppl": 2.376,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 140000,
      "training_tokens": 4587520000,
      "arc_easy_pct": 49.71,
      "blimp_pct": 77.58,
      "wikitext2_byte_ppl": 2.3639,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 160000,
      "training_tokens": 5242880000,
      "arc_easy_pct": 49.62,
      "blimp_pct": 77.14,
      "wikitext2_byte_ppl": 2.3523,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 180000,
      "training_tokens": 5898240000,
      "arc_easy_pct": 49.41,
      "blimp_pct": 77.65,
      "wikitext2_byte_ppl": 2.3455,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 200000,
      "training_tokens": 6553600000,
      "arc_easy_pct": 48.36,
      "blimp_pct": 78.65,
      "wikitext2_byte_ppl": 2.3406,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 220000,
      "training_tokens": 7208960000,
      "arc_easy_pct": 49.37,
      "blimp_pct": 78.53,
      "wikitext2_byte_ppl": 2.3275,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 240000,
      "training_tokens": 7864320000,
      "arc_easy_pct": 48.91,
      "blimp_pct": 77.67,
      "wikitext2_byte_ppl": 2.3287,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 260000,
      "training_tokens": 8519680000,
      "arc_easy_pct": 49.79,
      "blimp_pct": 78.53,
      "wikitext2_byte_ppl": 2.315,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 280000,
      "training_tokens": 9175040000,
      "arc_easy_pct": 50.17,
      "blimp_pct": 78.74,
      "wikitext2_byte_ppl": 2.3145,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 300000,
      "training_tokens": 9830400000,
      "arc_easy_pct": 50.46,
      "blimp_pct": 78.07,
      "wikitext2_byte_ppl": 2.3108,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 320000,
      "training_tokens": 10485760000,
      "arc_easy_pct": 50.46,
      "blimp_pct": 77.79,
      "wikitext2_byte_ppl": 2.3026,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 340000,
      "training_tokens": 11141120000,
      "arc_easy_pct": 50.13,
      "blimp_pct": 78.32,
      "wikitext2_byte_ppl": 2.304,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 360000,
      "training_tokens": 11796480000,
      "arc_easy_pct": 50.84,
      "blimp_pct": 78.01,
      "wikitext2_byte_ppl": 2.3011,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 380000,
      "training_tokens": 12451840000,
      "arc_easy_pct": 51.18,
      "blimp_pct": 78.86,
      "wikitext2_byte_ppl": 2.2994,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 400000,
      "training_tokens": 13107200000,
      "arc_easy_pct": 51.73,
      "blimp_pct": 79.38,
      "wikitext2_byte_ppl": 2.2966,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 420000,
      "training_tokens": 13762560000,
      "arc_easy_pct": 51.73,
      "blimp_pct": 79.26,
      "wikitext2_byte_ppl": 2.2919,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 440000,
      "training_tokens": 14417920000,
      "arc_easy_pct": 51.3,
      "blimp_pct": 78.63,
      "wikitext2_byte_ppl": 2.2835,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 460000,
      "training_tokens": 15073280000,
      "arc_easy_pct": 52.15,
      "blimp_pct": 78.6,
      "wikitext2_byte_ppl": 2.2793,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 480000,
      "training_tokens": 15728640000,
      "arc_easy_pct": 50.13,
      "blimp_pct": 78.66,
      "wikitext2_byte_ppl": 2.2777,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 500000,
      "training_tokens": 16384000000,
      "arc_easy_pct": 51.94,
      "blimp_pct": 78.16,
      "wikitext2_byte_ppl": 2.2758,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 520000,
      "training_tokens": 17039360000,
      "arc_easy_pct": 51.64,
      "blimp_pct": 78.93,
      "wikitext2_byte_ppl": 2.2718,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 540000,
      "training_tokens": 17694720000,
      "arc_easy_pct": 51.98,
      "blimp_pct": 78.67,
      "wikitext2_byte_ppl": 2.2684,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 560000,
      "training_tokens": 18350080000,
      "arc_easy_pct": 52.44,
      "blimp_pct": 78.79,
      "wikitext2_byte_ppl": 2.2692,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 580000,
      "training_tokens": 19005440000,
      "arc_easy_pct": 52.69,
      "blimp_pct": 79.43,
      "wikitext2_byte_ppl": 2.2628,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 600000,
      "training_tokens": 19660800000,
      "arc_easy_pct": 51.77,
      "blimp_pct": 79.32,
      "wikitext2_byte_ppl": 2.2625,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 620000,
      "training_tokens": 20316160000,
      "arc_easy_pct": 52.57,
      "blimp_pct": 79.31,
      "wikitext2_byte_ppl": 2.2585,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 640000,
      "training_tokens": 20971520000,
      "arc_easy_pct": 53.16,
      "blimp_pct": 79.49,
      "wikitext2_byte_ppl": 2.2578,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 660000,
      "training_tokens": 21626880000,
      "arc_easy_pct": 52.69,
      "blimp_pct": 79.59,
      "wikitext2_byte_ppl": 2.2587,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 680000,
      "training_tokens": 22282240000,
      "arc_easy_pct": 53.28,
      "blimp_pct": 79.28,
      "wikitext2_byte_ppl": 2.2562,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 700000,
      "training_tokens": 22937600000,
      "arc_easy_pct": 53.24,
      "blimp_pct": 78.98,
      "wikitext2_byte_ppl": 2.2563,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 720000,
      "training_tokens": 23592960000,
      "arc_easy_pct": 52.65,
      "blimp_pct": 78.92,
      "wikitext2_byte_ppl": 2.2523,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 740000,
      "training_tokens": 24248320000,
      "arc_easy_pct": 53.2,
      "blimp_pct": 79.48,
      "wikitext2_byte_ppl": 2.2533,
      "status": "intermediate"
    },
    {
      "model": "128M",
      "step": 760000,
      "training_tokens": 24903680000,
      "arc_easy_pct": 53.24,
      "blimp_pct": 79.09,
      "wikitext2_byte_ppl": 2.2538,
      "status": "final"
    }
  ]
}