Publish full native-grouped MMLU results and raise overview table

Move the results-at-a-glance table directly below the opening pitch and above the caveat. Fill the completed native-grouped 10% and 20% full-MMLU values, add ratio-labelled downstream rows, sync the compact summary, and publish the latest report.

README.md

@@ -25,6 +25,20 @@ identifies **10% of Laguna XS.2's routed MoE parameters** for removal while pres
 structure designed for efficient deployment kernels. That is **3.14B removable parameters**, or
 **9.39% of the full 33.44B-parameter model**.
 
+## Results at a glance
+
+| Native grouped pruning | Full-cache loss | Full MMLU 0-shot accuracy | MMLU-STEM 5-shot accuracy | GSM8K-CoT 8-shot strict EM |
+| --- | ---: | ---: | ---: | ---: |
+| BF16 baseline | 2.347142 | 0.733514 | 0.690771 | 0.931818 |
+| 10% pruning | 2.363612 | 0.735437 | 0.693942 | 0.909091 |
+| 20% pruning | 2.416442 | 0.734012 | 0.679036 | 0.780303 |
+
+Loss uses the full held-out 4k cache. GSM8K-CoT uses the paired first-10% subset (`132` examples).
+Full MMLU 0-shot remains effectively flat at 10% and 20%. Native-grouped MMLU-STEM 5-shot changes
+from `0.690771` to `0.693942` (`+0.003172`) at 10%, then to `0.679036` (`-0.011735`) at 20%.
+We treat the small positive deltas as retained quality, not evidence that pruning improves the
+model. The separate atomic-pruning MMLU sweep is not substituted here.
+
 > [!IMPORTANT]
 > The current artifact measures pruning quality, not deployment speed. It applies the structured
 > pruning mask by zeroing blocks inside the original tensors, so the released Laguna kernel still
@@ -38,21 +52,6 @@ structured removal: `3,140,616,192` parameters, or `9.39%` of the full model. Fu
 perplexity moves from `10.455647` to `10.629271` (`+1.66%`), while strict GSM8K-CoT exact match on
 a paired first-10% subset moves from `0.931818` to `0.909091`.
 
-## Results at a glance
-
-| Native grouped pruning | Full-cache loss | Full MMLU 0-shot accuracy | MMLU-STEM 5-shot accuracy | GSM8K-CoT 8-shot strict EM |
-| --- | ---: | ---: | ---: | ---: |
-| BF16 baseline | 2.347142 | 0.733514 | 0.690771 | 0.931818 |
-| 10% pruning | 2.363612 | not run | 0.693942 | 0.909091 |
-| 20% pruning | 2.416442 | not run | 0.679036 | 0.780303 |
-
-Loss uses the full held-out 4k cache. GSM8K-CoT uses the paired first-10% subset (`132` examples).
-Native-grouped full MMLU 0-shot was run for the baseline and the exploratory 25% stress point.
-Native-grouped MMLU-STEM 5-shot was also run for the 10% and 20% masks. Accuracy changes from
-`0.690771` to `0.693942` (`+0.003172`) at 10%, then to `0.679036` (`-0.011735`) at 20%. We treat
-the small positive 10% delta as retained quality, not evidence that pruning improves the model.
-The separate atomic-pruning MMLU sweep is not substituted here.
-
 ## What structured pruning means here
 
 Laguna routes tokens to parent experts. Inside each parent expert, the computation can be decomposed
@@ -215,7 +214,9 @@ more sensitive:
 
 | Benchmark | BF16 baseline | Structured-pruning variant | Delta |
 | --- | ---: | ---: | ---: |
-| Full MMLU 0-shot accuracy | 0.733514 | 0.725965 | -0.007549 |
+| Full MMLU 0-shot accuracy, 10% pruning | 0.733514 | 0.735437 | +0.001923 |
+| Full MMLU 0-shot accuracy, 20% pruning | 0.733514 | 0.734012 | +0.000499 |
+| Full MMLU 0-shot accuracy, 25% pruning | 0.733514 | 0.725965 | -0.007549 |
 | MMLU-STEM 5-shot accuracy, 10% pruning | 0.690771 | 0.693942 | +0.003172 |
 | MMLU-STEM 5-shot accuracy, 20% pruning | 0.690771 | 0.679036 | -0.011735 |
 | MMLU-STEM 5-shot accuracy, 25% pruning | 0.690771 | 0.666350 | -0.024421 |
@@ -305,7 +306,7 @@ uv run --extra dev python scripts/plot_pruning_insights.py
 ## Status
 
 - Completed: native grouped loss sweep at 10%, 20%, 25%, and 40%.
-- Completed: full baseline and native-grouped 25% MMLU.
+- Completed: full baseline and native-grouped 10%, 20%, and 25% MMLU.
 - Completed: exploratory repacked child-routing loss sweep.
 - Completed: paired first-10% GSM8K-CoT baseline and native-grouped 10%, 20%, and 25% curve.
 - Stopped after smoke validation: paired CRUXEval-O CoT comparison, due to the remaining time box.

reports/pruning_eval_progress.md

@@ -26,6 +26,13 @@ router. It zeroes importance-ranked 64-atom blocks only inside routed expert ten
 GSM8K-CoT values use the same first 10% subset (`132` examples), fixed eight-shot prompt, and greedy
 decoding for every row.
 
+| Native grouped pruning | Full-cache loss | Full MMLU 0-shot accuracy | MMLU-STEM 5-shot accuracy | GSM8K-CoT 8-shot strict EM |
+| --- | ---: | ---: | ---: | ---: |
+| BF16 baseline | 2.347142 | 0.733514 | 0.690771 | 0.931818 |
+| 10% | 2.363612 | 0.735437 | 0.693942 | 0.909091 |
+| 20% | 2.416442 | 0.734012 | 0.679036 | 0.780303 |
+| 25% | 2.466706 | 0.725965 | 0.666350 | 0.681818 |
+
 ## Loss / Perplexity
 
 | Run | Pruning interpretation | Mean loss | Delta loss | Perplexity | PPL delta | Artifact |
@@ -78,6 +85,8 @@ Per-layer 25% mask statistics: `artifacts/reports/native_group_pruning_stats.md`
 | Full MMLU 0-shot atomic 10% | full 14,042 samples | accuracy 0.737715; delta +0.004202 | `artifacts/runs/downstream/mmlu_atomic_10pct_full.json` |
 | Full MMLU 0-shot atomic 20% | full 14,042 samples | accuracy 0.734511; delta +0.000997 | `artifacts/runs/downstream/mmlu_atomic_20pct_full.json` |
 | Full MMLU 0-shot atomic 40% | full 14,042 samples | accuracy 0.681028; delta -0.052485 | `artifacts/runs/downstream/mmlu_atomic_40pct_full.json` |
+| Full MMLU 0-shot native grouped 10% | full 14,042 samples | accuracy 0.735437; delta +0.001923 | `artifacts/runs/downstream/mmlu_native_group_10pct_full.json` |
+| Full MMLU 0-shot native grouped 20% | full 14,042 samples | accuracy 0.734012; delta +0.000499 | `artifacts/runs/downstream/mmlu_native_group_20pct_full.json` |
 | Full MMLU 0-shot native grouped 25% | full 14,042 samples | accuracy 0.725965; delta -0.007549; STEM slice 0.688233 | `artifacts/runs/downstream/mmlu_native_group_25pct_full.json` |
 | MMLU-STEM 5-shot baseline | report-metric-aligned local lm-eval; 3,153 samples | accuracy 0.690771 | `artifacts/runs/downstream/mmlu_stem_5shot_baseline_full.json` |
 | MMLU-STEM 5-shot native grouped 10% | report-metric-aligned local lm-eval; 3,153 samples | accuracy 0.693942; delta +0.003172 | `artifacts/runs/downstream/mmlu_stem_5shot_native_group_10pct_full.json` |

results/summary.json

@@ -172,7 +172,14 @@
     },
     "mmlu": {
       "num_examples": 14042,
-      "num_fewshot": 0
+      "num_fewshot": 0,
+      "baseline_accuracy": 0.733514,
+      "native_group_10pct_accuracy": 0.735437,
+      "native_group_10pct_delta": 0.001923,
+      "native_group_20pct_accuracy": 0.734012,
+      "native_group_20pct_delta": 0.000499,
+      "native_group_25pct_accuracy": 0.725965,
+      "native_group_25pct_delta": -0.007549
     },
     "mmlu_stem_5shot": {
       "num_examples": 3153,