maxgfr commited on
Commit
64238b0
·
verified ·
1 Parent(s): a323603

Publish s6-micro-v1

Browse files
Files changed (1) hide show
  1. README.md +2 -2
README.md CHANGED
@@ -144,8 +144,8 @@ Face, classic datasets converted into typed questions, game states labelled by e
144
  numeric curriculum, the same as s6-nano. Pretraining text: SimpleStories, TinyStories, FineWeb and
145
  the non-web parts of Dolma 3 and TxT360. Soft labels on 747,077 training questions come from
146
  GLM-5.3-flash (open weights, MIT, called through the z.ai API), kept only where they agree with
147
- the gold answer; no test question was sent. The z.ai terms of service restrict training on its
148
- outputs; the project's choice and the labels are described in
149
  [DATA_LICENSES.md](https://github.com/maxgfr/skycmd/blob/main/DATA_LICENSES.md), with every
150
  other license.
151
 
 
144
  numeric curriculum, the same as s6-nano. Pretraining text: SimpleStories, TinyStories, FineWeb and
145
  the non-web parts of Dolma 3 and TxT360. Soft labels on 747,077 training questions come from
146
  GLM-5.3-flash (open weights, MIT, called through the z.ai API), kept only where they agree with
147
+ the gold answer; no test question was sent. The teacher's weights are open (MIT); the labels and
148
+ their publication are described in
149
  [DATA_LICENSES.md](https://github.com/maxgfr/skycmd/blob/main/DATA_LICENSES.md), with every
150
  other license.
151