Add model accuracy and local runtime comparison charts

.gitattributes

@@ -40,3 +40,5 @@ figures/arrmes_v7_open_model_accuracy_comparison.png filter=lfs diff=lfs merge=l
 figures/arrmes_v7_runtime_latency_memory_footprint.png filter=lfs diff=lfs merge=lfs -text
 figures/arrmes_v7_track_c_workflow_scorecard.png filter=lfs diff=lfs merge=lfs -text
 tokenizer.json filter=lfs diff=lfs merge=lfs -text
+figures/arrmes-model-accuracy-comparison.png filter=lfs diff=lfs merge=lfs -text
+figures/arrmes-local-runtime-comparison.png filter=lfs diff=lfs merge=lfs -text

README.md

@@ -24,7 +24,9 @@ tags:
 - **Subdirectory `v7-lora/` (`v1.4.0`):** Contains the `v7` LoRA adapter trained on `TrainPool-v7` (`5,827` examples, `seed=20261102`, `200` iterations, `lr=1e-4`) and its provenance manifest (`v7-lora/provenance.json`).
 - **Gemma Terms of Use Notice:** Distribution and use of these weights are subject to the [Gemma Terms of Use](https://ai.google.dev/gemma/terms) and include [`GEMMA_TERMS_NOTICE.txt`](GEMMA_TERMS_NOTICE.txt) per Section 3.1.
 
-![Arrmes v1.4.0 Locked Acceptance-v5 Accuracy and Confidence Bounds](arrmes_v7_open_model_accuracy_comparison.png)
+![Raw and verified accuracy for five Arrmes generative checkpoints on the same 200-case acceptance-v5 slice](figures/arrmes-model-accuracy-comparison.png)
+
+Five generative checkpoints on the same 100 positive and 100 negative cases; raw model output and verified proposals are shown separately. This slice is separate from the full-suite neural-router results. [Recorded scores](https://github.com/spencernorton/arrmes/blob/bcac7c2fb83b4e6c263ac804245433b249e4d48f/experiments/acceptance-v5/summary.json).
 
 ## Evaluation Results (Paired Base vs Fine-Tuned & Locked `acceptance-v5`)
 
@@ -35,12 +37,14 @@ Raw model output (`raw_unverified`, Track A), verifier-processed proposals (`raw
   - **Root Public `v5` Adapter (`96e89a7...`):** **`63 / 80` (`78.8%`) raw exact** (`+51` calls), **`68 / 80` (`85.0%`) after verification** (`+45` calls), and **`29 / 35` strict fixture workflows** (`0` unexpected write proposals). One normalized template in the 80-case diagnostic overlaps training (`ON1-023`), so it is recorded as a known-input diagnostic rather than a template-disjoint acceptance suite.
 - **Locked `acceptance-v5` Stratified 200-Case Generative Comparison (`seed=7007`, [`NorviTech/arrmes-bench-v5`](https://github.com/spencernorton/arrmes)):**
   - Evaluated on a stratified 200-case slice (`100` positive + `100` hard-negative/safety) with **0 exact query overlaps** and **0 entity overlaps** against `TrainPool-v7` (`5,827` rows) and `acceptance-v4` (`1,000` rows):
-    - **Public `FunctionGemma-270M v5` (`bf16 + root LoRA`)**: Positive Track A **`68 / 100` (`68.0%`)**, Track B **`78 / 100` (`78.0%`)**; Hard-Negative Track A `33 / 100`, Track B **`79 / 100` (`79.0%`)**, **`0` unauthorized writes**.
-    - **Trained `FunctionGemma-270M v7` (`4-bit MLX` fused from `v7-lora/`, `175.83 MiB`)**: Positive Track A **`70 / 100` (`70.0%`)**, Track B **`91 / 100` (`91.0%`, `+13.0 pp` over `v5`)**; Hard-Negative Track A `35 / 100`, Track B **`88 / 100` (`88.0%`, `+9.0 pp` over `v5`)**, **`0` unauthorized writes**.
+    - **Public `FunctionGemma-270M v5` (`bf16 + root LoRA`)**: Positive Track A **`68 / 100` (`68.0%`)**, Track B **`78 / 100` (`78.0%`)**; Hard-Negative Track A `30 / 100`, Track B **`87 / 100` (`87.0%`)**, **`0` unauthorized writes**.
+    - **Trained `FunctionGemma-270M v7` (`4-bit MLX` fused from `v7-lora/`, `175.83 MiB`)**: Positive Track A **`70 / 100` (`70.0%`)**, Track B **`91 / 100` (`91.0%`, `+13.0 pp` over `v5`)**; Hard-Negative Track A `35 / 100`, Track B **`88 / 100` (`88.0%`, `+1.0 pp` over `v5`)**, **`0` unauthorized writes**.
 - **Historical `sealed-acceptance-v3` (`N = 314`) Audit Note:**
   - The 4,389-row historical training file shares 37 normalized templates with 261/314 `sealed-v3` rows (`226` non-write opportunities). Use [`NorviTech/arrmes-bench-v5`](https://github.com/spencernorton/arrmes) for zero-leakage evaluation, and [`NorviTech/arrmes-neural-router-v3`](https://huggingface.co/NorviTech/arrmes-neural-router-v3) (`7.52 MiB`) when `>=99%` raw Track A accuracy and `arrmes.tools.v3` coverage are required.
 
-![Cross-Platform Runtime Latency and Memory Footprint](arrmes_v7_claude_sonnet_and_resource_comparison.png)
+![Warm routing latency and model size for the neural router, FunctionGemma and Qwen3 on Apple M4 Pro](figures/arrmes-local-runtime-comparison.png)
+
+Warm routing plus verification on Apple M4 Pro, excluding service and network time. Disk size is separate from runtime memory; sample counts and axis scales are labeled. [Recorded profiles](https://github.com/spencernorton/arrmes/blob/bcac7c2fb83b4e6c263ac804245433b249e4d48f/experiments/runtime-profile-v7.json).
 
 ## Measured Runtime Profile (Apple M4 Pro, `v1.4.0`)