correction: Atlas gain is not statistically significant; label KAGL/Polyvore as significant

README.md

@@ -35,7 +35,7 @@ model-index:
 
 # MODA Pro Lite
 
-**A 213M open-weights fashion encoder that beats [MODA](https://huggingface.co/HopitAI/moda-fashionsiglip-multiview-203m) — the strongest open fashion retrieval system — on catalog search, at full corpus, from one checkpoint with no serving machinery.**
+**A 213M open-weights fashion encoder that beats [MODA](https://huggingface.co/HopitAI/moda-fashionsiglip-multiview-203m) — the strongest open fashion retrieval system — on catalog search, at full corpus and with statistical significance, from one checkpoint with no serving machinery.**
 
 By [Hopit AI](https://hopit.ai). Weights are open (Apache-2.0); the training and data pipeline are proprietary.
 
@@ -43,10 +43,9 @@ By [Hopit AI](https://hopit.ai). Weights are open (Apache-2.0); the training and
 |---|--:|--:|
 | KAGL (44K products) | **0.3185** vs 0.2890 | **+10.2%** |
 | Polyvore (94K) | **0.3997** vs 0.3726 | **+7.3%** |
-| Atlas (78K) | **0.1945** vs 0.1884 | **+3.2%** |
 | ZooClaw-Fashion short | **0.4994** vs 0.4775 | **+4.6%** |
 
-Every number in this card is a **full-corpus** measurement (no subsampled galleries) through one shared harness, alongside FashionSigLIP, SigLIP-SO400M, and ZooClaw under identical protocol. Full six-system table: [benchmark page](https://hopit-ai.github.io/Moda/).
+The KAGL and Polyvore gains are significant under a paired bootstrap (10,000 resamples, 95% CI excludes zero); on Atlas the difference is positive but inconclusive, and on Fashion200K and In-Shop MODA is significantly better — see the full table. Every number in this card is a **full-corpus** measurement (no subsampled galleries) through one shared harness, alongside FashionSigLIP, SigLIP-SO400M, and ZooClaw under identical protocol. Full six-system table: [benchmark page](https://hopit-ai.github.io/Moda/).
 
 ## The MODA family