# Model configuration
model_path: utils/qwen2.5_1.5b_curr_step_400  # Change to the smaller model
max_tokens: 1024  # From max_response_length in training config
temperature: 0.7  # Lower temperature for more focused responses
top_p: 0.9  # From rollout top_p
developer_prompt: DeepSeekZero
developer_role: system  # Standard role for system prompts

# Output configuration
output_dir: eval_results
save_metadata: true
save_full_results: true
eval_repeats: 3

# Categories and datasets to evaluate
categories:
  - category: reasoning
    datasets:
      - dataset: spell_backward
        size: 100
        seed: 42
        params:
          min_word_len: 3  # From training config
          max_word_len: 10
          data_file: holdout_words.txt