[Eval] Random walk on a 2D grid

[elh]

Thank you for contributing an eval! ♥️

🚨 Please make sure your PR follows these guidelines, failure to follow the guidelines below will result in the PR being closed automatically. Note that even if the criteria are met, that does not guarantee the PR will be merged nor GPT-4 access granted. 🚨

PLEASE READ THIS:

In order for a PR to be merged, it must fail on GPT-4. We are aware that right now, users do not have access, so you will not be able to tell if the eval fails or not. Please run your eval with GPT-3.5-Turbo, but keep in mind as we run the eval, if GPT-4 gets higher than 90% on the eval, we will likely reject since GPT-4 is already capable of completing the task.

We plan to roll out a way for users submitting evals to see the eval performance on GPT-4 soon. Stay tuned! Until then, you will not be able to see the eval performance on GPT-4. We encourage partial PR's with ~5-10 example that we can then run the evals on and share the results with you so you know how your eval does with GPT-4 before writing all 100 examples.

Eval details 📑

Eval name

2d_movement

Eval description

Evaluate if GPT can keep track of its position and orientation while random walking on a 2D grid in first-person POV

What makes this a useful eval?

This is a simple-to-describe, trivial-for-humans, test of spatial reasoning. This feels like a core problem for natural language control of an agent in a 2D space: physical or virtual.

I think this is a good candidate for "physical reasoning" in the "Building an eval" category list. The samples include varying lengths of steps but GPT fails even very short ones.

🎯 9% accuracy on gpt-3.5-turbo and 20% accuracy on gpt-4

Criteria for a good eval ✅

Below are some of the criteria we look for in a good eval. In general, we are seeking cases where the model does not do a good job despite being capable of generating a good response (note that there are some things large language models cannot do, so those would not make good evals).

Your eval should be:

• Thematically consistent: The eval should be thematically consistent. We'd like to see a number of prompts all demonstrating some particular failure mode. For example, we can create an eval on cases where the model fails to reason about the physical world.
• Contains failures where a human can do the task, but either GPT-4 or GPT-3.5-Turbo could not.
• Includes good signal around what is the right behavior. This means either a correct answer for Basic evals or the Fact Model-graded eval, or an exhaustive rubric for evaluating answers for the Criteria Model-graded eval.
• Include at least 100 high quality examples (it is okay to only contribute 5-10 meaningful examples and have us test them with GPT-4 before adding all 100)

If there is anything else that makes your eval worth including, please document it below.

Unique eval value

Insert what makes your eval high quality that was not mentioned above. (Not required)

Eval structure 🏗️

Your eval should

• Check that your data is in evals/registry/data/{name}
• Check that your yaml is registered at evals/registry/evals/{name}.yaml
• Ensure you have the right to use the data you submit via this eval

(For now, we will only be approving evals that use one of the existing eval classes. You may still write custom eval classes for your own cases, and we may consider merging them in the future.)

Final checklist 👀

Submission agreement

By contributing to Evals, you are agreeing to make your evaluation logic and data under the same MIT license as this repository. You must have adequate rights to upload any data used in an Eval. OpenAI reserves the right to use this data in future service improvements to our product. Contributions to OpenAI Evals will be subject to our usual Usage Policies (https://platform.openai.com/docs/usage-policies).

• I agree that my submission will be made available under an MIT license and complies with OpenAI's usage policies.

Email address validation

If your submission is accepted, we will be granting GPT-4 access to a limited number of contributors. Access will be given to the email address associated with the merged pull request.

• I acknowledge that GPT-4 access will only be granted, if applicable, to the email address used for my merged pull request.

Limited availability acknowledgement

We know that you might be excited to contribute to OpenAI's mission, help improve our models, and gain access to GPT-4. However, due to the requirements mentioned above and high volume of submissions, we will not be able to accept all submissions and thus not grant everyone who opens a PR GPT-4 access. We know this is disappointing, but we hope to set the right expectation before you open this PR.

• I understand that opening a PR, even if it meets the requirements above, does not guarantee the PR will be merged nor GPT-4 access granted.

Submit eval

• I have filled out all required fields in the evals PR form
• (Ignore if not submitting code) I have run pip install pre-commit; pre-commit install and have verified that black, isort, and autoflake are running when I commit and push

Failure to fill out all required fields will result in the PR being closed.

Eval JSON data

Since we are using Git LFS, we are asking eval submitters to add in as many Eval Samples (at least 5) from their contribution here:

View evals in JSON

Eval

{"input": [{"role": "system", "content": "Please note: In the following EXERCISE, it is important that you only respond with a single line in the format (x, y). Imagine you are standing in a 2D coordinate grid at (0, 0) where coordinates are represented like (x, y). You are currently facing the positive y direction."}, {"role": "user", "content": "EXERCISE: If you take 5 steps forward, turn left, take 2 steps forward, turn left, take 1 step backward, turn left, take two steps backward, what coordinate are you at?"}], "ideal": "(-4, 6)"}
{"input": [{"role": "system", "content": "Please note: In the following EXERCISE, it is important that you only respond with a single line in the format (x, y). Imagine you are standing in a 2D coordinate grid at (0, 0) where coordinates are represented like (x, y). You are currently facing the positive y direction."}, {"role": "user", "content": "EXERCISE: If you turn 90 degrees left, then take 3 steps forward, what coordinate are you at?"}], "ideal": "(-3, 0)"}
{"input": [{"role": "system", "content": "Please note: In the following EXERCISE, it is important that you only respond with a single line in the format (x, y). Imagine you are standing in a 2D coordinate grid at (0, 0) where coordinates are represented like (x, y). You are currently facing the positive y direction."}, {"role": "user", "content": "EXERCISE: If you turn 90 degrees right, then turn 90 degrees left, then take 4 steps backward, then take 5 steps backward, what coordinate are you at?"}], "ideal": "(0, -9)"}
{"input": [{"role": "system", "content": "Please note: In the following EXERCISE, it is important that you only respond with a single line in the format (x, y). Imagine you are standing in a 2D coordinate grid at (0, 0) where coordinates are represented like (x, y). You are currently facing the positive y direction."}, {"role": "user", "content": "EXERCISE: If you take 5 steps forward, then take 2 steps forward, then take 5 steps backward, then turn 90 degrees right, then take 2 steps forward, then take 3 steps forward, then turn 90 degrees left, then turn 90 degrees right, then take 1 step backward, then turn 90 degrees left, what coordinate are you at?"}], "ideal": "(4, 2)"}
{"input": [{"role": "system", "content": "Please note: In the following EXERCISE, it is important that you only respond with a single line in the format (x, y). Imagine you are standing in a 2D coordinate grid at (0, 0) where coordinates are represented like (x, y). You are currently facing the positive y direction."}, {"role": "user", "content": "EXERCISE: If you take 2 steps backward, then take 1 step backward, then turn 90 degrees left, then take 4 steps forward, then turn 90 degrees left, then turn 90 degrees left, then take 1 step forward, then turn 90 degrees left, then turn 90 degrees left, then take 2 steps forward, then turn 90 degrees right, then turn 90 degrees right, what coordinate are you at?"}], "ideal": "(-5, -3)"}

[elh]

Samples generated via https://github.com/elh/make-evals/blob/main/2d_movement/main.py

[elh]

🎯 9% accuracy on gpt-3.5-turbo and 20% accuracy on gpt-4

[sushant-openai]

"Thanks for opening this PR. It seems to be a good eval.

I would like to recommend some changes in the eval's model. Currently, you are using Match.
But in the instructions for the model you have mentioned the model to respond in a single line in the format (x, y).
So here the model may take its flexibility to respond with some extra explanation too.

In such case, we can simply check if the Ideal completion is in the model's response or not. This can be done using the Includes.
Also Could you please add a proper description in the evals/registry/evals/2d_movement.yaml.

We would love to review your PR again after you make this change."

[sushant-openai]

"Thanks for opening this PR. It seems to be a good eval.

I would like to recommend some changes in the eval's model. Currently, you are using Match.
But in the instructions for the model you have mentioned the model to respond in a single line in the format (x, y).
So here the model may take its flexibility to respond with some extra explanation too.

In such case, we can simply check if the Ideal completion is in the model's response or not. This can be done using the Includes.
Also Could you please add a proper description in the evals/registry/evals/2d_movement.yaml.

We would love to review your PR again after you make this change."

[elh]

• description added
  • minor doc update to help let people know this is expected [docs] document that description is an expected eval field in build-eval.md #1014
• Use Includes instead of Match. Same general performance. 🎯 9% accuracy on gpt-3.5-turbo and 20% accuracy on gpt-4
  • When I wrote my eval, Includes was buggy. Glad to see it has been fixed. Let a few open PRs and Issues know that it has been patched here Fix get_answer #972

Think this is ready for review again @sushant-openai

[sushant-openai]

Thanks for making the requested PR changes. I'm approving this eval.