v8: continued training to 2.6B tokens (from 1.51B). Improved coherence on narrative prompts. Same architecture (10,052,864 params, 5L d320 5-head SwiGLU RoPE). (#14)

- v8: continued training to 2.6B tokens (from 1.51B). Improved coherence on narrative prompts. Same architecture (10,052,864 params, 5L d320 5-head SwiGLU RoPE). (adec7f2dee715ec4253caeceef3db40df34e3e23)

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e5b543e3b0b13dd9444c47d0f264c7b6f97c5f3bea9e7747565eae774188cd87
-size 41142264
+oid sha256:5c3ce3c6c4ac2561c5be3c27f3a12a7a8272338cfee74e29d02bb8826bb6f798
+size 41142232