Card: BLiMP count note — cause of the 47 913 count not determined

README.md

@@ -29,7 +29,7 @@ weights via `load_slayer.py`, and the trainer checkpoint directly); both gave th
 
 | task | metric | n | score |
 |---|---|---:|---:|
-| BLiMP (67 phenomena) | accuracy | 67 000 | 71.51 % (47 912 correct; one evaluation on a different GPU counted 47 913, one pair flipping at a numerical tie) |
+| BLiMP (67 phenomena) | accuracy | 67 000 | 71.51 % (47 912 correct; an evaluation on another GPU counted 47 913, cause not determined: no per-item record from that run) |
 | ARC-Easy (test) | accuracy | 2 376 | 42.89 % (1 019 correct) |
 | WikiText-2 (raw, test) | token perplexity (this tokenizer) | — | 49.9734 |
 | WikiText-2 (raw, test) | byte perplexity | 1 292 008 bytes | 2.5857 |