LittlePlyModel + harness

CM-13 / final snapshot

CM-13: clean full-move history.

Historical source checkpoint at update 279,393.

Fresh random initialization with AdamW; clean game history, one token per complete UCI move (one ply), including promotions. The 86,966,016-parameter transformer has 12 layers, 12 heads and width 768, with a 512-token context and 1,975 vocabulary entries. No masking and no outcome labels in inputs. Move outputs and fixed-White WIN/DRAW/LOSS outputs are separately normalized from the shared transformer and tied output matrix.

Fresh random initialization with AdamW; no parent checkpoint. Source 2db161c6c8396388565f414123ccbe413deec2d0a85faee530ea8b5913a16ae4; inference 745b457f7751d912b878fd5242243e058ad2da2fe3eae1b002fbc40d2ffd0cbc.

Matched Sunfish benchmarkWins / draws / lossesScore
CM-13 final, 100 games63 / 9 / 2867.5%
CM-13 eight-hour reference, 100 games47 / 12 / 4153%

Observed gain: 14.5 percentage points. Paired opening-cluster bootstrap 95% interval: +1.5 to +27 percentage points (50 paired openings, 20,000 resamples). Draws count as half a point. This supports improvement on this Sunfish benchmark; it is not general Elo or a browser-strength estimate.

Full experiment · Eight-hour snapshot