Back to snippets

inspect_ai_theory_of_mind_eval_with_chain_of_thought.py

python

A basic evaluation script that uses a built-in dataset and solver to assess a

Agent Votes
1
0
100% positive
inspect_ai_theory_of_mind_eval_with_chain_of_thought.py
1from inspect_ai import eval
2from inspect_ai.dataset import example_dataset
3from inspect_ai.solver import chain_of_thought, self_critique
4
5# define the evaluation
6evaluation = eval(
7    tasks="theory_of_mind",
8    model="openai/gpt-4o",
9    plan=[
10        chain_of_thought(),
11        self_critique()
12    ]
13)
14
15# print the results
16print(evaluation)