Historical material
Preserved research material.
Earlier formats, artifacts and limitations remain available; current entry points are in the main navigation.
Archived notebook content
Historical CM-14 checkpoint
The 86,966,016-parameter transformer reads clean, complete move history: one token per UCI move, including promotions. It has 12 layers, 12 heads, width 768, 512 context positions and 1,975 vocabulary entries. Move logits and fixed-White win/draw/loss logits are normalized separately. Outcome labels are not supplied as input, and the WDL diagnostic does not choose the browser move.
CM-14 began from fresh random initialization with a hybrid Muon/AdamW optimizer. This promoted checkpoint continues that CM-14 run; it is not a continuation of CM-13 weights.
Historical notebook material and artifacts
Current final snapshot
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.
Current input and serving interfaces · Final evaluation · Published browser package. Earlier explanations below retain their historical context; CM-10 is no longer the browser default.
Earlier notebook material and artifacts
Online on Lichess: CM-13, audited eight-hour snapshot. Play LittlePly on Lichess, with a different tokenizer and helper. CM-13 results and limitations.
Current online model: CM-13
CM-13 starts from random initialization with AdamW. Clean game history is encoded as one token per complete UCI move (one ply), including promotions, in a 512-token context. There is no masking and no outcome label in the inputs.
A shared 12-layer, 768-wide, 12-head transformer and tied output matrix produce two separately normalized distributions: move outputs and fixed-White WIN/DRAW/LOSS outputs. Outcome predictions are diagnostics, not policy conditioning or a search value function.
The model selects ordinary legal moves. A separate asynchronous Stockfish 19 helper can intervene for explicit offensive or defensive mates; centipawn scores do not trigger changes. This deployed system is not search-free. Engine mate reports are not independently verified proof certificates.
Eight-hour checkpoint, results and training context; serving metadata. The CM-10 explanations and artifacts below describe the separate browser model and earlier experiment.
Learning guide
A move is three tokens.
The model predicts a move from examples. Chess rules and external search are separate parts of the playing system.
Represent a move
A ply is one player's move. Squares a1–h8 map to IDs 0–63. Each move contributes source square, destination square and promotion-or-none; 64 means no promotion. CM-10 begins with BOS 69 and retains the complete move history.
Predict and select
A causal transformer predicts the next token from earlier tokens. For a candidate move, the destination depends on the history and source; promotion depends on the history, source and destination. Legal candidates are ranked by their joint likelihood. The external mate helper is a separate search and does not imply a model reasoning trace.
Read the evidence
CM-10 checkpoint · Evaluation and limitations · Retired browser artifact · Reproduction instructions
Compare input families
Historical CM-6/CM-7 instead use a state prefix and 512 vocabulary IDs; see the historical state model source. Compare exact input formats. Losses from different input/target objectives are not directly comparable.