From 6617353b8457201b2b38483fdb58d470bc331c1a Mon Sep 17 00:00:00 2001 From: Charles Martin Date: Sat, 29 Aug 2026 22:08:01 -0700 Subject: [PATCH] Move four-head nanoGPT results to dedicated experiment --- .../README.md | 10 ++++++++++ .../results/five_seed}/plots/adamw_D.png | Bin .../results/five_seed}/plots/adamw_ERG_gap.png | Bin .../five_seed}/plots/adamw_alpha_clip_xmax.png | Bin .../plots/adamw_alpha_clip_xmax_zoomed.png | Bin .../results/five_seed}/plots/adamw_alpha_raw.png | Bin .../five_seed}/plots/adamw_alpha_raw_zoomed.png | Bin .../results/five_seed}/plots/adamw_num_traps.png | Bin .../five_seed}/plots/adamw_rand_distance.png | Bin .../five_seed}/plots/grad_norm_pre_clip.png | Bin .../results/five_seed}/plots/muon_clip_D.png | Bin .../results/five_seed}/plots/muon_clip_ERG_gap.png | Bin .../five_seed}/plots/muon_clip_alpha_clip_xmax.png | Bin .../plots/muon_clip_alpha_clip_xmax_zoomed.png | Bin .../five_seed}/plots/muon_clip_alpha_raw.png | Bin .../plots/muon_clip_alpha_raw_zoomed.png | Bin .../five_seed}/plots/muon_clip_num_traps.png | Bin .../five_seed}/plots/muon_clip_rand_distance.png | Bin .../results/five_seed}/plots/train_loss.png | Bin .../five_seed}/plots/update_to_weight_ratio.png | Bin .../results/five_seed}/plots/val_accuracy.png | Bin .../five_seed}/plots/val_generalization_gap.png | Bin .../results/five_seed}/plots/val_loss.png | Bin .../results/five_seed}/plots/val_perplexity.png | Bin .../results/five_seed}/plots/weight_norm.png | Bin .../results/five_seed}/protocol.yaml | 0 .../results/five_seed}/report.html | 0 .../results/README.md | 13 +++++++------ 28 files changed, 17 insertions(+), 6 deletions(-) create mode 100644 baseline/experiments/nanogpt_four_head_2026_08_27_baseline/README.md rename baseline/experiments/{nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed => nanogpt_four_head_2026_08_27_baseline/results/five_seed}/plots/adamw_D.png (100%) rename baseline/experiments/{nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed => nanogpt_four_head_2026_08_27_baseline/results/five_seed}/plots/adamw_ERG_gap.png (100%) rename baseline/experiments/{nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed => nanogpt_four_head_2026_08_27_baseline/results/five_seed}/plots/adamw_alpha_clip_xmax.png (100%) rename baseline/experiments/{nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed => nanogpt_four_head_2026_08_27_baseline/results/five_seed}/plots/adamw_alpha_clip_xmax_zoomed.png (100%) rename baseline/experiments/{nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed => nanogpt_four_head_2026_08_27_baseline/results/five_seed}/plots/adamw_alpha_raw.png (100%) rename baseline/experiments/{nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed => nanogpt_four_head_2026_08_27_baseline/results/five_seed}/plots/adamw_alpha_raw_zoomed.png (100%) rename baseline/experiments/{nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed => nanogpt_four_head_2026_08_27_baseline/results/five_seed}/plots/adamw_num_traps.png (100%) rename baseline/experiments/{nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed => nanogpt_four_head_2026_08_27_baseline/results/five_seed}/plots/adamw_rand_distance.png (100%) rename baseline/experiments/{nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed => nanogpt_four_head_2026_08_27_baseline/results/five_seed}/plots/grad_norm_pre_clip.png (100%) rename baseline/experiments/{nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed => nanogpt_four_head_2026_08_27_baseline/results/five_seed}/plots/muon_clip_D.png (100%) rename baseline/experiments/{nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed => nanogpt_four_head_2026_08_27_baseline/results/five_seed}/plots/muon_clip_ERG_gap.png (100%) rename baseline/experiments/{nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed => nanogpt_four_head_2026_08_27_baseline/results/five_seed}/plots/muon_clip_alpha_clip_xmax.png (100%) rename baseline/experiments/{nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed => nanogpt_four_head_2026_08_27_baseline/results/five_seed}/plots/muon_clip_alpha_clip_xmax_zoomed.png (100%) rename baseline/experiments/{nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed => nanogpt_four_head_2026_08_27_baseline/results/five_seed}/plots/muon_clip_alpha_raw.png (100%) rename baseline/experiments/{nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed => nanogpt_four_head_2026_08_27_baseline/results/five_seed}/plots/muon_clip_alpha_raw_zoomed.png (100%) rename baseline/experiments/{nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed => nanogpt_four_head_2026_08_27_baseline/results/five_seed}/plots/muon_clip_num_traps.png (100%) rename baseline/experiments/{nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed => nanogpt_four_head_2026_08_27_baseline/results/five_seed}/plots/muon_clip_rand_distance.png (100%) rename baseline/experiments/{nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed => nanogpt_four_head_2026_08_27_baseline/results/five_seed}/plots/train_loss.png (100%) rename baseline/experiments/{nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed => nanogpt_four_head_2026_08_27_baseline/results/five_seed}/plots/update_to_weight_ratio.png (100%) rename baseline/experiments/{nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed => nanogpt_four_head_2026_08_27_baseline/results/five_seed}/plots/val_accuracy.png (100%) rename baseline/experiments/{nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed => nanogpt_four_head_2026_08_27_baseline/results/five_seed}/plots/val_generalization_gap.png (100%) rename baseline/experiments/{nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed => nanogpt_four_head_2026_08_27_baseline/results/five_seed}/plots/val_loss.png (100%) rename baseline/experiments/{nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed => nanogpt_four_head_2026_08_27_baseline/results/five_seed}/plots/val_perplexity.png (100%) rename baseline/experiments/{nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed => nanogpt_four_head_2026_08_27_baseline/results/five_seed}/plots/weight_norm.png (100%) rename baseline/experiments/{nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed => nanogpt_four_head_2026_08_27_baseline/results/five_seed}/protocol.yaml (100%) rename baseline/experiments/{nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed => nanogpt_four_head_2026_08_27_baseline/results/five_seed}/report.html (100%) diff --git a/baseline/experiments/nanogpt_four_head_2026_08_27_baseline/README.md b/baseline/experiments/nanogpt_four_head_2026_08_27_baseline/README.md new file mode 100644 index 0000000..33b3bad --- /dev/null +++ b/baseline/experiments/nanogpt_four_head_2026_08_27_baseline/README.md @@ -0,0 +1,10 @@ +# Four-head nanoGPT baseline — 2026-08-27 + +This experiment archives the completed five-seed AdamW versus MuonClip four-attention-head baseline. + +- [Open rendered report](https://raw.githack.com/CalculatedContent/rg_optimizers/main/baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/report.html) +- [Archived HTML](results/five_seed/report.html) +- [Exact protocol](results/five_seed/protocol.yaml) +- [Plot assets](results/five_seed/plots/) + +The campaign used seeds 1337, 2027, 4099, 31415, and 271828. The report contains five-seed aggregate performance, WeightWatcher, and optimizer-diagnostic plots. diff --git a/baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/adamw_D.png b/baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/adamw_D.png similarity index 100% rename from baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/adamw_D.png rename to baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/adamw_D.png diff --git a/baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/adamw_ERG_gap.png b/baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/adamw_ERG_gap.png similarity index 100% rename from baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/adamw_ERG_gap.png rename to baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/adamw_ERG_gap.png diff --git a/baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/adamw_alpha_clip_xmax.png b/baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/adamw_alpha_clip_xmax.png similarity index 100% rename from baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/adamw_alpha_clip_xmax.png rename to baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/adamw_alpha_clip_xmax.png diff --git a/baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/adamw_alpha_clip_xmax_zoomed.png b/baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/adamw_alpha_clip_xmax_zoomed.png similarity index 100% rename from baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/adamw_alpha_clip_xmax_zoomed.png rename to baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/adamw_alpha_clip_xmax_zoomed.png diff --git a/baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/adamw_alpha_raw.png b/baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/adamw_alpha_raw.png similarity index 100% rename from baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/adamw_alpha_raw.png rename to baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/adamw_alpha_raw.png diff --git a/baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/adamw_alpha_raw_zoomed.png b/baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/adamw_alpha_raw_zoomed.png similarity index 100% rename from baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/adamw_alpha_raw_zoomed.png rename to baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/adamw_alpha_raw_zoomed.png diff --git a/baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/adamw_num_traps.png b/baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/adamw_num_traps.png similarity index 100% rename from baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/adamw_num_traps.png rename to baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/adamw_num_traps.png diff --git a/baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/adamw_rand_distance.png b/baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/adamw_rand_distance.png similarity index 100% rename from baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/adamw_rand_distance.png rename to baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/adamw_rand_distance.png diff --git a/baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/grad_norm_pre_clip.png b/baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/grad_norm_pre_clip.png similarity index 100% rename from baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/grad_norm_pre_clip.png rename to baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/grad_norm_pre_clip.png diff --git a/baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/muon_clip_D.png b/baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/muon_clip_D.png similarity index 100% rename from baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/muon_clip_D.png rename to baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/muon_clip_D.png diff --git a/baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/muon_clip_ERG_gap.png b/baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/muon_clip_ERG_gap.png similarity index 100% rename from baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/muon_clip_ERG_gap.png rename to baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/muon_clip_ERG_gap.png diff --git a/baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/muon_clip_alpha_clip_xmax.png b/baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/muon_clip_alpha_clip_xmax.png similarity index 100% rename from baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/muon_clip_alpha_clip_xmax.png rename to baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/muon_clip_alpha_clip_xmax.png diff --git a/baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/muon_clip_alpha_clip_xmax_zoomed.png b/baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/muon_clip_alpha_clip_xmax_zoomed.png similarity index 100% rename from baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/muon_clip_alpha_clip_xmax_zoomed.png rename to baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/muon_clip_alpha_clip_xmax_zoomed.png diff --git a/baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/muon_clip_alpha_raw.png b/baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/muon_clip_alpha_raw.png similarity index 100% rename from baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/muon_clip_alpha_raw.png rename to baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/muon_clip_alpha_raw.png diff --git a/baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/muon_clip_alpha_raw_zoomed.png b/baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/muon_clip_alpha_raw_zoomed.png similarity index 100% rename from baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/muon_clip_alpha_raw_zoomed.png rename to baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/muon_clip_alpha_raw_zoomed.png diff --git a/baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/muon_clip_num_traps.png b/baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/muon_clip_num_traps.png similarity index 100% rename from baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/muon_clip_num_traps.png rename to baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/muon_clip_num_traps.png diff --git a/baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/muon_clip_rand_distance.png b/baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/muon_clip_rand_distance.png similarity index 100% rename from baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/muon_clip_rand_distance.png rename to baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/muon_clip_rand_distance.png diff --git a/baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/train_loss.png b/baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/train_loss.png similarity index 100% rename from baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/train_loss.png rename to baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/train_loss.png diff --git a/baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/update_to_weight_ratio.png b/baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/update_to_weight_ratio.png similarity index 100% rename from baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/update_to_weight_ratio.png rename to baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/update_to_weight_ratio.png diff --git a/baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/val_accuracy.png b/baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/val_accuracy.png similarity index 100% rename from baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/val_accuracy.png rename to baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/val_accuracy.png diff --git a/baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/val_generalization_gap.png b/baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/val_generalization_gap.png similarity index 100% rename from baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/val_generalization_gap.png rename to baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/val_generalization_gap.png diff --git a/baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/val_loss.png b/baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/val_loss.png similarity index 100% rename from baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/val_loss.png rename to baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/val_loss.png diff --git a/baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/val_perplexity.png b/baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/val_perplexity.png similarity index 100% rename from baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/val_perplexity.png rename to baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/val_perplexity.png diff --git a/baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/weight_norm.png b/baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/weight_norm.png similarity index 100% rename from baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/plots/weight_norm.png rename to baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/plots/weight_norm.png diff --git a/baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/protocol.yaml b/baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/protocol.yaml similarity index 100% rename from baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/protocol.yaml rename to baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/protocol.yaml diff --git a/baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/report.html b/baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/report.html similarity index 100% rename from baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/report.html rename to baseline/experiments/nanogpt_four_head_2026_08_27_baseline/results/five_seed/report.html diff --git a/baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/README.md b/baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/README.md index c2663db..097a944 100644 --- a/baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/README.md +++ b/baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/README.md @@ -1,8 +1,9 @@ -# Archived nanoGPT baseline results +# One-head nanoGPT baseline results -These reports archive the completed five-seed AdamW and MuonClip baselines. The rendered links are pinned to commit `19e3c280d55beb89cdf68e23288dab78ccb0607d`, including every referenced plot asset. +This directory contains the completed five-seed AdamW and MuonClip one-head baseline. -| Experiment | Rendered report | Archived files | -|---|---|---| -| One attention head, five seeds | [Open rendered report](https://raw.githack.com/CalculatedContent/rg_optimizers/19e3c280d55beb89cdf68e23288dab78ccb0607d/baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/one_head_5seed/report.html) | [HTML](one_head_5seed/report.html) · [Protocol](one_head_5seed/protocol.yaml) | -| Four attention heads, five seeds | [Open rendered report](https://raw.githack.com/CalculatedContent/rg_optimizers/19e3c280d55beb89cdf68e23288dab78ccb0607d/baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/four_head_5seed/report.html) | [HTML](four_head_5seed/report.html) · [Protocol](four_head_5seed/protocol.yaml) | +- [Open rendered report](https://raw.githack.com/CalculatedContent/rg_optimizers/main/baseline/experiments/nanogpt_one_head_2026_08_21_baseline/results/one_head_5seed/report.html) +- [Archived HTML](one_head_5seed/report.html) +- [Exact protocol](one_head_5seed/protocol.yaml) + +The four-head baseline is archived separately under [`nanogpt_four_head_2026_08_27_baseline`](../../nanogpt_four_head_2026_08_27_baseline/).