e4b_pure_opd_v6a_norm025

When

Recorded date: 2026-11-02

First archive commit: 2026-11-04T17:35:03-05:00 (offset shown).

Training run day: 2026-11-02 (America/New_York). For derived variants, this is the parent run day, not the variant’s creation day.

First Git archive commit time (UTC offset shown), not the time training changed weights. Per-update wall timestamps were not preserved.

Exact tensor-change timestamp: unavailable.

Why change it?

Test a smaller adapter perturbation after the v6a training run.

Which part changed?

Adapter matrices over the frozen base, targeting: ^model\.language_model\.layers\.\d+\.(self_attn\.(q_proj|k_proj|v_proj|o_proj)|mlp\.(gate_proj|up_proj|down_proj))$

The source describes adapter targets and mathematical changes. No before/after tensor comparison is available in this viewer.

How was it changed?

Post-training norm-adjusted v6a variant, labelled norm025 in the historical ledger; the base run used pure online prediction distillation. Exported low-rank adapter, rank 64, scaling alpha 32.

Did it improve?

Historical measurements · Historical local result for norm025: 11/45. A selected variant, not a confirmed held-out improvement.

Evaluation set: Historical sets vary; see source table · Batch: Not applicable. Local screening and historical task scores are not competition scores unless explicitly identified in the source. Different evaluation sets cannot be compared as a gain.

Original evidence

• v2/experiments/adapters/e4b_pure_opd_v6a_norm025/checkpoint_manifest.json
• v2/experiments/adapters/e4b_pure_opd_v6a_norm025/adapter_config.json
• docs/v2-post-training-and-lb-postmortem.md

Literal source excerpts follow. They may contain original predictions and superseded counts; use the measured-outcome summary and corrections above for the current result.

{
  "base_model_name_or_path": "gemma-4-e4b-it-qat-w4a16-ct",
  "bias": "none",
  "lora_alpha": 32,
  "lora_dropout": 0.0,
  "peft_type": "LORA",
  "r": 64,
  "target_modules": "^model\\.language_model\\.layers\\.\\d+\\.(self_attn\\.(q_proj|k_proj|v_proj|o_proj)|mlp\\.(gate_proj|up_proj|down_proj))$",
  "task_type": "CAUSAL_LM",
  "use_rslora": false
}

{
  "adapter_id": "e4b_pure_opd_v6a_norm025",
  "wrx90_path": "/home/keith/gemma-v2/adapters/e4b_pure_opd_v6a_norm025",
  "files": {
    "adapter_model.safetensors": {
      "size_bytes": 279129832,
      "sha256": "14006e5d4a653c7cd7938258dec6c25d9a0e9d81bafe699b1527741954cb3066"
    }
  }
}