Add disclaimers and clarify dataset purpose in README

Restore disclaimer that PAWN is under active development by a single
developer. Clarify that datasets are for adapter training (behavioral
cloning), not pretraining — PAWN itself is trained only on random games.

Author: thomas-schweich

README.md

@@ -32,7 +32,9 @@ I've found PAWN to be a viable testbed for finetuning and augmentation methods a
 
 Finetuning PAWN has proven significantly more parameter-efficient than training new models from scratch and requires minimal compute resources.
 
-Feel free to use PAWN in your own experiments. If you spot a bug, please help out by creating an issue or PR.
+Feel free to use PAWN in your own experiments. Note that PAWN was developed as a personal project by a single developer and has not been published or audited. If you spot a bug, please help out by creating an issue or PR.
+
+**PAWN is under active development and is not yet stable.**
 
 ## Model Variants
 
@@ -103,6 +105,8 @@ uv run python -m pawn.dashboard --log-dir logs  # real-time monitoring
 
 ## Datasets
 
+These datasets are for **adapter training (behavioral cloning)**, not for pretraining PAWN itself. PAWN is pretrained exclusively on random legal games generated on-the-fly -- it never sees human or engine games during pretraining. The datasets below provide real gameplay data for finetuning the frozen PAWN backbone into player models that mimic specific playstyles or skill levels.
+
 | Dataset | Games | Description | Link |
 |---------|-------|-------------|------|
 | Lichess Full | ~289M train + 50K val + 50K test | Rated games from Q1 2025 (all Elos), holdout from Jan 2026 | [pawn-lichess-full](https://huggingface.co/datasets/thomas-schweich/pawn-lichess-full) |