tdelab/tde-general-v0.2

TDE general v0.2 — the typed-decision encoder, now also playing Snake

The same 150M encoder as tde-general-v0.1 (ModernBERT-base, pointer readout; one forward pass over state + question + candidates returns a probability per candidate), fine-tuned further so it plays Snake from a plain-text board while keeping its general decisions.

Snake

The state is the board as text, one token per cell ( . empty, F food, H head, 1– 9 body: moves until that cell is free) under a line with the length and heading; the question is "Which way should the snake move?" with the candidates up, down, left and right. The top candidate is the move. An anti-trap shield vetoes a move after which the head can no longer reach its tail and takes the model's next choice instead.

24x16 board, initial length 6, seeds 101–104, 600 moves per game	food per game	survived
tde-general-v0.2 with the anti-trap shield	39 · 38 · 40 · 38 (mean 38.75)	4 of 4
tde-general-v0.2, model alone	27 · 18 · 13 · 29 (mean 21.75)	0 of 4

About 12 ms per move with MLX on an M5 Pro. A live page where the model plays game after game: python -m integrations.snake.demo in the code repository (tutorial: integrations/snake).

General decisions

In-distribution test (7 public datasets, 2,000-item sample)	accuracy	NLL
tde-general-v0.1	87.4%	0.313
tde-general-v0.2	86.4%	0.336

JevBench and the typed-decisions test were not re-run for this checkpoint.

Training

Initialised from tde-general-v0.1 and fine-tuned for 2 epochs (6,250 steps of 32) on 100k Snake positions from 8x8, 12x12 and 24x16 boards, each labelled with the moves of a shortest-path planner that keeps its tail reachable. Trained with MLX on one Mac in about 80 minutes (integrations/snake/build_sft.py, then tde/mlx/train.py).

Usage

# pip install git+https://github.com/e13ven-arch/tde   (package name: tde)
from tde.inference import Decider
d = Decider.from_run("tdelab/tde-general-v0.2")   # or a local release folder
d.decide(board_text, {"type": "choice", "instructions": "Which way should the snake move?",
                      "criteria": {m: m for m in ("up", "down", "left", "right")}})

General questions work exactly as with v0.1.

Limitations

Without the shield the snake eventually traps itself: all four protocol games end before move 600. General accuracy is about one point below v0.1; the limitations listed for v0.1 apply unchanged.