Fix over-strong 'no token loops' claim: 40-sample sweep (seeds 0-4) found 1 hard loop (seed 4, loop_frac 1.0) + several elevated-loop samples. Card now says 'occasional/rare token loops' instead of 'no token loops'. All other numbers (val 3.8943, ppl 49.12, ~308M tok) re-verified and unchanged.

README.md

@@ -59,7 +59,7 @@ Standard LLaMA block: RMSNorm → MHA (RoPE) → residual → RMSNorm → SwiGLU
 
 DedeProGames requested **2.6B tokens**. This checkpoint is at **~308M tokens** — an **8.4× shortfall**. The GPU was occupied by other work for most of the training window.
 
-At 30 tok/param the model has learned the **surface shape** of English much better than v1 — real words, parseable sentences, no token loops — but the prose is still **semantically incoherent** (word salad): grammatically plausible sentences that don't mean what they say. The val loss (3.89) is well below the 7.38 unigram floor, so it genuinely uses context; the improvement from 4.60 → 3.89 is real.
+At 30 tok/param the model has learned the **surface shape** of English much better than v1 — real words, parseable sentences, occasional token loops — but the prose is still **semantically incoherent** (word salad): grammatically plausible sentences that don't mean what they say. The val loss (3.89) is well below the 7.38 unigram floor, so it genuinely uses context; the improvement from 4.60 → 3.89 is real.
 
 This is a **continued checkpoint**, not the final deliverable. More data is the fix; continued training toward the 2.6B budget is planned.
 
@@ -70,7 +70,7 @@ This is a **continued checkpoint**, not the final deliverable. More data is the
 | val loss | 4.6020 | **3.8943** |
 | perplexity | 99.68 | **49.12** |
 | Below unigram floor (7.38)? | Yes | Yes |
-| Token loops? | No | No |
+| Token loops? | No | Rare (1 hard loop in a 40-sample sweep, seeds 0-4) |
 | Semantically coherent? | No | No (improved, still word salad) |
 
 ### Sample outputs (v4, real generation from the weights)
@@ -83,7 +83,7 @@ This is a **continued checkpoint**, not the final deliverable. More data is the
 
 > "To be or not to be is a question about the world. The last thing about the world is a part of a real life. It would be a great place, like it and in the earth."
 
-These are real outputs from the v4 weights (not hand-picked for coherence). The text is grammatically structured — real words, parseable sentences, no token loops, no broken tokens — but semantically incoherent. That is the honest state of a 10M model at 30 tok/param.
+These are real outputs from the v4 weights (not hand-picked for coherence). The text is grammatically structured — real words, parseable sentences, occasional token loops, no broken tokens — but semantically incoherent. That is the honest state of a 10M model at 30 tok/param.
 
 ## Usage