Upload mixture final checkpoint (step 1421)

README.md

@@ -33,7 +33,7 @@ the **final-step checkpoint** (end of the single training epoch).
 | seed | 42 |
 
 Validation loss at the final checkpoint (mean cross-entropy / scored token):
-- `val_mix`: 1.7961
+- `val_mix`: 1.7983
 
 ## Usage
 
@@ -51,5 +51,7 @@ out = model.generate(seed, max_new_tokens=150, do_sample=True, temperature=1.0,
 print(tokenizer.decode(out[0][1:], skip_special_tokens=True))
 ```
 
-Tokenization convention used in training: `add_special_tokens=False`, EOS (id=1)
-appended to every story, truncated to 512 tokens.
+Tokenization convention used in training: `add_special_tokens=False`, every story
+wrapped in EOS (id=1) on both sides — `[EOS, tokens…, EOS]` — truncated to 512
+tokens. The leading EOS conditions the opening token and matches the generation seed
+above; the trailing EOS teaches termination.

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8706b53e9dabb4ae241effa2fb212794a6a7c1c2af41058a9e469426df4b07ca
+oid sha256:fd3a828a4d863c4be9c3c82a800cd08acc4dc2c82d5f8561abcf7c4929ea88cd
 size 21424032