Added Icelandic inflection eval; JsonMatch eval function #1387
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
vthorsteinsson
requested review from
andrew-openai,
jwang47 and
logankilpatrick
as code owners
|
|
||
| result = self.completion_fn( | ||
| prompt=prompt, | ||
| temperature=0.0, # Q: why are these hardcoded? |
There was a problem hiding this comment.
We set temp=0 for determinism in running the eval, otherwise results won't necessarily be replicable.
If there's a clear need for nonzero temp in the eval, we can definitely add it as an arg!
There was a problem hiding this comment.
Yes, understood - this comment and question was in the original FuzzyMatch code that I used as a template for JsonMatch, so I just left it in ;-)
|
The Icelandic Eval and JSON Match look great, thanks for adding these, approving+merging. |
andrew-openai
approved these changes
Oct 27, 2023
jacobbieker
pushed a commit
to withmartian/-ARCHIVED--router-evals
that referenced
this pull request
Jan 9, 2024
# 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 be 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 it 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. **Starting April 10, the minimum eval count is 15 samples, we hope this makes it easier to create and contribute evals.** Also, please note that we're using **Git LFS** for storing the JSON files, so please make sure that you move the JSON file to Git LFS before submitting a PR. Details on how to use Git LFS are available [here](https://git-lfs.com). ## Eval details 📑 ### Eval name Icelandic noun phrase inflection ### Eval description This eval consists of 3 x 100 samples in "easy", "medium" and "hard" categories. Each sample represents the task of inflecting a noun phrase in Icelandic, in all four cases of the language (nominative, accusative, dative and genitive), both singular and plural. A noun phrase consists of an adjective and a noun (e.g., "fallegur litur" = "beautiful color"). In the easy category, both the adjective and the noun are relatively common. In the medium category, they are less common, and in the hard category they are rare enough that it is pretty unlikely that they occur in any training corpora. ### What makes this a useful eval? The eval is designed to test the general grammatical proficiency of a model in Icelandic, and the eval accuracy is assumed to correlate with a model's ability to generate grammatically correct text in the language. GPT models have so far struggled with generating correct Icelandic text, even though GPT-4 was uniquely trained by RLHF in the language. Icelandic is believed to be a good bellwether for lower-resource, grammatically complex language support in general. Inflecting noun phrases is something that native language speakers do without significant effort, even if they have not seen the particular adjective and the noun before, as it can be done on the basis of generic grammatical pattern recognition. However, to date, GPT-4 seems not to have acquired enough of a "native feel" for Icelandic to be able to do this task with high accuracy. ## 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: - [x] 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. - [x] Contains failures where a human can do the task, but either GPT-4 or GPT-3.5-Turbo could not. - [x] 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. - [x] **Include at least 15 high-quality examples.** 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 - [x] Check that your data is in `evals/registry/data/{name}` - [x] Check that your YAML is registered at `evals/registry/evals/{name}.yaml` - [x] 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.) **Note: this PR includes a new general eval class, JsonMatch, which is not specific to the Icelandic evaluation case. It allows completions and ideal answers to be represented as JSON objects, comparing the objects by individual key:value pairs. Tests and documentation of this functionality are included in the PR.** ## 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>). - [x] 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 commits on the merged pull request. - [x] I acknowledge that GPT-4 access will only be granted, if applicable, to the email address used for my merged pull request. ### Limited availability acknowledgment 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 the 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. - [x] 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 be granted. ### Submit eval - [x] I have filled out all required fields of this form - [x] I have used **Git LFS** for the Eval JSON data - [x] (Ignore if not submitting code) I have run `pip install pre-commit; pre-commit install` and have verified that `mypy`, `black`, `isort`, `autoflake` and `ruff` 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: <details> <summary>View evals in JSON</summary> ### Eval ```jsonl {"input": [{"role": "system", "content": "Þú ert sérfræðingur í íslenskri málfræði."}, {"role": "user", "content": "Hvernig fallbeygist nafnliðurinn \"palestínskur fréttavefur\" í öllum föllum (nf, þf, þgf, ef), eintölu (et) og fleirtölu (ft), án greinis? Svaraðu í *JSON formi eingöngu* og auðkenndu tölur og föll með skammstöfunum et, ft, nf, þf, þgf, ef."}], "ideal": "{\"et\": {\"nf\": \"palestínskur fréttavefur\", \"þf\": \"palestínskan fréttavef\", \"þgf\": \"palestínskum fréttavef\", \"ef\": \"palestínsks fréttavefjar\"}, \"ft\": {\"nf\": \"palestínskir fréttavefir\", \"þf\": \"palestínska fréttavefi\", \"þgf\": \"palestínskum fréttavefjum\", \"ef\": \"palestínskra fréttavefja\"}}"} {"input": [{"role": "system", "content": "Þú ert sérfræðingur í íslenskri málfræði."}, {"role": "user", "content": "Hvernig fallbeygist nafnliðurinn \"hliðhollt lyfjapróf\" í öllum föllum (nf, þf, þgf, ef), eintölu (et) og fleirtölu (ft), án greinis? Svaraðu í *JSON formi eingöngu* og auðkenndu tölur og föll með skammstöfunum et, ft, nf, þf, þgf, ef."}], "ideal": "{\"et\": {\"nf\": \"hliðhollt lyfjapróf\", \"þf\": \"hliðhollt lyfjapróf\", \"þgf\": \"hliðhollu lyfjaprófi\", \"ef\": \"hliðholls lyfjaprófs\"}, \"ft\": {\"nf\": \"hliðholl lyfjapróf\", \"þf\": \"hliðholl lyfjapróf\", \"þgf\": \"hliðhollum lyfjaprófum\", \"ef\": \"hliðhollra lyfjaprófa\"}}"} {"input": [{"role": "system", "content": "Þú ert sérfræðingur í íslenskri málfræði."}, {"role": "user", "content": "Hvernig fallbeygist nafnliðurinn \"refsiverð stjörnuleit\" í öllum föllum (nf, þf, þgf, ef), eintölu (et) og fleirtölu (ft), án greinis? Svaraðu í *JSON formi eingöngu* og auðkenndu tölur og föll með skammstöfunum et, ft, nf, þf, þgf, ef."}], "ideal": "{\"et\": {\"nf\": \"refsiverð stjörnuleit\", \"þf\": \"refsiverða stjörnuleit\", \"þgf\": \"refsiverðri stjörnuleit\", \"ef\": \"refsiverðrar stjörnuleitar\"}, \"ft\": {\"nf\": \"refsiverðar stjörnuleitir\", \"þf\": \"refsiverðar stjörnuleitir\", \"þgf\": \"refsiverðum stjörnuleitum\", \"ef\": \"refsiverðra stjörnuleita\"}}"} {"input": [{"role": "system", "content": "Þú ert sérfræðingur í íslenskri málfræði."}, {"role": "user", "content": "Hvernig fallbeygist nafnliðurinn \"japönsk landbúnaðarvara\" í öllum föllum (nf, þf, þgf, ef), eintölu (et) og fleirtölu (ft), án greinis? Svaraðu í *JSON formi eingöngu* og auðkenndu tölur og föll með skammstöfunum et, ft, nf, þf, þgf, ef."}], "ideal": "{\"et\": {\"nf\": \"japönsk landbúnaðarvara\", \"þf\": \"japanska landbúnaðarvöru\", \"þgf\": \"japanskri landbúnaðarvöru\", \"ef\": \"japanskrar landbúnaðarvöru\"}, \"ft\": {\"nf\": \"japanskar landbúnaðarvörur\", \"þf\": \"japanskar landbúnaðarvörur\", \"þgf\": \"japönskum landbúnaðarvörum\", \"ef\": \"japanskra landbúnaðarvara\"}}"} {"input": [{"role": "system", "content": "Þú ert sérfræðingur í íslenskri málfræði."}, {"role": "user", "content": "Hvernig fallbeygist nafnliðurinn \"dýrmætt vistheimili\" í öllum föllum (nf, þf, þgf, ef), eintölu (et) og fleirtölu (ft), án greinis? Svaraðu í *JSON formi eingöngu* og auðkenndu tölur og föll með skammstöfunum et, ft, nf, þf, þgf, ef."}], "ideal": "{\"et\": {\"nf\": \"dýrmætt vistheimili\", \"þf\": \"dýrmætt vistheimili\", \"þgf\": \"dýrmætu vistheimili\", \"ef\": \"dýrmæts vistheimilis\"}, \"ft\": {\"nf\": \"dýrmæt vistheimili\", \"þf\": \"dýrmæt vistheimili\", \"þgf\": \"dýrmætum vistheimilum\", \"ef\": \"dýrmætra vistheimila\"}}"} ``` </details>
Linmj-Judy
pushed a commit
to TablewareBox/evals
that referenced
this pull request
Feb 27, 2024
# 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 be 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 it 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. **Starting April 10, the minimum eval count is 15 samples, we hope this makes it easier to create and contribute evals.** Also, please note that we're using **Git LFS** for storing the JSON files, so please make sure that you move the JSON file to Git LFS before submitting a PR. Details on how to use Git LFS are available [here](https://git-lfs.com). ## Eval details 📑 ### Eval name Icelandic noun phrase inflection ### Eval description This eval consists of 3 x 100 samples in "easy", "medium" and "hard" categories. Each sample represents the task of inflecting a noun phrase in Icelandic, in all four cases of the language (nominative, accusative, dative and genitive), both singular and plural. A noun phrase consists of an adjective and a noun (e.g., "fallegur litur" = "beautiful color"). In the easy category, both the adjective and the noun are relatively common. In the medium category, they are less common, and in the hard category they are rare enough that it is pretty unlikely that they occur in any training corpora. ### What makes this a useful eval? The eval is designed to test the general grammatical proficiency of a model in Icelandic, and the eval accuracy is assumed to correlate with a model's ability to generate grammatically correct text in the language. GPT models have so far struggled with generating correct Icelandic text, even though GPT-4 was uniquely trained by RLHF in the language. Icelandic is believed to be a good bellwether for lower-resource, grammatically complex language support in general. Inflecting noun phrases is something that native language speakers do without significant effort, even if they have not seen the particular adjective and the noun before, as it can be done on the basis of generic grammatical pattern recognition. However, to date, GPT-4 seems not to have acquired enough of a "native feel" for Icelandic to be able to do this task with high accuracy. ## 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: - [x] 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. - [x] Contains failures where a human can do the task, but either GPT-4 or GPT-3.5-Turbo could not. - [x] 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. - [x] **Include at least 15 high-quality examples.** 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 - [x] Check that your data is in `evals/registry/data/{name}` - [x] Check that your YAML is registered at `evals/registry/evals/{name}.yaml` - [x] 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.) **Note: this PR includes a new general eval class, JsonMatch, which is not specific to the Icelandic evaluation case. It allows completions and ideal answers to be represented as JSON objects, comparing the objects by individual key:value pairs. Tests and documentation of this functionality are included in the PR.** ## 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>). - [x] 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 commits on the merged pull request. - [x] I acknowledge that GPT-4 access will only be granted, if applicable, to the email address used for my merged pull request. ### Limited availability acknowledgment 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 the 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. - [x] 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 be granted. ### Submit eval - [x] I have filled out all required fields of this form - [x] I have used **Git LFS** for the Eval JSON data - [x] (Ignore if not submitting code) I have run `pip install pre-commit; pre-commit install` and have verified that `mypy`, `black`, `isort`, `autoflake` and `ruff` 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: <details> <summary>View evals in JSON</summary> ### Eval ```jsonl {"input": [{"role": "system", "content": "Þú ert sérfræðingur í íslenskri málfræði."}, {"role": "user", "content": "Hvernig fallbeygist nafnliðurinn \"palestínskur fréttavefur\" í öllum föllum (nf, þf, þgf, ef), eintölu (et) og fleirtölu (ft), án greinis? Svaraðu í *JSON formi eingöngu* og auðkenndu tölur og föll með skammstöfunum et, ft, nf, þf, þgf, ef."}], "ideal": "{\"et\": {\"nf\": \"palestínskur fréttavefur\", \"þf\": \"palestínskan fréttavef\", \"þgf\": \"palestínskum fréttavef\", \"ef\": \"palestínsks fréttavefjar\"}, \"ft\": {\"nf\": \"palestínskir fréttavefir\", \"þf\": \"palestínska fréttavefi\", \"þgf\": \"palestínskum fréttavefjum\", \"ef\": \"palestínskra fréttavefja\"}}"} {"input": [{"role": "system", "content": "Þú ert sérfræðingur í íslenskri málfræði."}, {"role": "user", "content": "Hvernig fallbeygist nafnliðurinn \"hliðhollt lyfjapróf\" í öllum föllum (nf, þf, þgf, ef), eintölu (et) og fleirtölu (ft), án greinis? Svaraðu í *JSON formi eingöngu* og auðkenndu tölur og föll með skammstöfunum et, ft, nf, þf, þgf, ef."}], "ideal": "{\"et\": {\"nf\": \"hliðhollt lyfjapróf\", \"þf\": \"hliðhollt lyfjapróf\", \"þgf\": \"hliðhollu lyfjaprófi\", \"ef\": \"hliðholls lyfjaprófs\"}, \"ft\": {\"nf\": \"hliðholl lyfjapróf\", \"þf\": \"hliðholl lyfjapróf\", \"þgf\": \"hliðhollum lyfjaprófum\", \"ef\": \"hliðhollra lyfjaprófa\"}}"} {"input": [{"role": "system", "content": "Þú ert sérfræðingur í íslenskri málfræði."}, {"role": "user", "content": "Hvernig fallbeygist nafnliðurinn \"refsiverð stjörnuleit\" í öllum föllum (nf, þf, þgf, ef), eintölu (et) og fleirtölu (ft), án greinis? Svaraðu í *JSON formi eingöngu* og auðkenndu tölur og föll með skammstöfunum et, ft, nf, þf, þgf, ef."}], "ideal": "{\"et\": {\"nf\": \"refsiverð stjörnuleit\", \"þf\": \"refsiverða stjörnuleit\", \"þgf\": \"refsiverðri stjörnuleit\", \"ef\": \"refsiverðrar stjörnuleitar\"}, \"ft\": {\"nf\": \"refsiverðar stjörnuleitir\", \"þf\": \"refsiverðar stjörnuleitir\", \"þgf\": \"refsiverðum stjörnuleitum\", \"ef\": \"refsiverðra stjörnuleita\"}}"} {"input": [{"role": "system", "content": "Þú ert sérfræðingur í íslenskri málfræði."}, {"role": "user", "content": "Hvernig fallbeygist nafnliðurinn \"japönsk landbúnaðarvara\" í öllum föllum (nf, þf, þgf, ef), eintölu (et) og fleirtölu (ft), án greinis? Svaraðu í *JSON formi eingöngu* og auðkenndu tölur og föll með skammstöfunum et, ft, nf, þf, þgf, ef."}], "ideal": "{\"et\": {\"nf\": \"japönsk landbúnaðarvara\", \"þf\": \"japanska landbúnaðarvöru\", \"þgf\": \"japanskri landbúnaðarvöru\", \"ef\": \"japanskrar landbúnaðarvöru\"}, \"ft\": {\"nf\": \"japanskar landbúnaðarvörur\", \"þf\": \"japanskar landbúnaðarvörur\", \"þgf\": \"japönskum landbúnaðarvörum\", \"ef\": \"japanskra landbúnaðarvara\"}}"} {"input": [{"role": "system", "content": "Þú ert sérfræðingur í íslenskri málfræði."}, {"role": "user", "content": "Hvernig fallbeygist nafnliðurinn \"dýrmætt vistheimili\" í öllum föllum (nf, þf, þgf, ef), eintölu (et) og fleirtölu (ft), án greinis? Svaraðu í *JSON formi eingöngu* og auðkenndu tölur og föll með skammstöfunum et, ft, nf, þf, þgf, ef."}], "ideal": "{\"et\": {\"nf\": \"dýrmætt vistheimili\", \"þf\": \"dýrmætt vistheimili\", \"þgf\": \"dýrmætu vistheimili\", \"ef\": \"dýrmæts vistheimilis\"}, \"ft\": {\"nf\": \"dýrmæt vistheimili\", \"þf\": \"dýrmæt vistheimili\", \"þgf\": \"dýrmætum vistheimilum\", \"ef\": \"dýrmætra vistheimila\"}}"} ``` </details>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
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 be 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 it 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. Starting April 10, the minimum eval count is 15 samples, we hope this makes it easier to create and contribute evals.
Also, please note that we're using Git LFS for storing the JSON files, so please make sure that you move the JSON file to Git LFS before submitting a PR. Details on how to use Git LFS are available here.
Eval details 📑
Eval name
Icelandic noun phrase inflection
Eval description
This eval consists of 3 x 100 samples in "easy", "medium" and "hard" categories. Each sample
represents the task of inflecting a noun phrase in Icelandic, in all four cases of the language
(nominative, accusative, dative and genitive), both singular and plural. A noun phrase
consists of an adjective and a noun (e.g., "fallegur litur" = "beautiful color").
In the easy category, both the adjective and the noun are
relatively common. In the medium category, they are less common, and in the hard category they
are rare enough that it is pretty unlikely that they occur in any training corpora.
What makes this a useful eval?
The eval is designed to test the general grammatical proficiency of a model in Icelandic, and
the eval accuracy is assumed to correlate with a model's ability to generate grammatically
correct text in the language. GPT models have so far struggled with generating correct Icelandic
text, even though GPT-4 was uniquely trained by RLHF in the language. Icelandic is believed to
be a good bellwether for lower-resource, grammatically complex language support in general.
Inflecting noun phrases is something that native language speakers do without significant
effort, even if they have not seen the particular adjective and the noun before, as it can be done on the
basis of generic grammatical pattern recognition. However, to date, GPT-4 seems not to have
acquired enough of a "native feel" for Icelandic to be able to do this task with high accuracy.
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:
Basicevals or theFactModel-graded eval, or an exhaustive rubric for evaluating answers for theCriteriaModel-graded eval.If there is anything else that makes your eval worth including, please document it below.
Unique eval value
Eval structure 🏗️
Your eval should
evals/registry/data/{name}evals/registry/evals/{name}.yaml(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.)
Note: this PR includes a new general eval class, JsonMatch, which is not specific to the Icelandic evaluation
case. It allows completions and ideal answers to be represented as JSON objects, comparing the objects
by individual key:value pairs. Tests and documentation of this functionality are included in the PR.
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).
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 commits on the merged pull request.
Limited availability acknowledgment
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 the 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.
Submit eval
pip install pre-commit; pre-commit installand have verified thatmypy,black,isort,autoflakeandruffare running when I commit and pushFailure 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": "Þú ert sérfræðingur í íslenskri málfræði."}, {"role": "user", "content": "Hvernig fallbeygist nafnliðurinn \"palestínskur fréttavefur\" í öllum föllum (nf, þf, þgf, ef), eintölu (et) og fleirtölu (ft), án greinis? Svaraðu í *JSON formi eingöngu* og auðkenndu tölur og föll með skammstöfunum et, ft, nf, þf, þgf, ef."}], "ideal": "{\"et\": {\"nf\": \"palestínskur fréttavefur\", \"þf\": \"palestínskan fréttavef\", \"þgf\": \"palestínskum fréttavef\", \"ef\": \"palestínsks fréttavefjar\"}, \"ft\": {\"nf\": \"palestínskir fréttavefir\", \"þf\": \"palestínska fréttavefi\", \"þgf\": \"palestínskum fréttavefjum\", \"ef\": \"palestínskra fréttavefja\"}}"} {"input": [{"role": "system", "content": "Þú ert sérfræðingur í íslenskri málfræði."}, {"role": "user", "content": "Hvernig fallbeygist nafnliðurinn \"hliðhollt lyfjapróf\" í öllum föllum (nf, þf, þgf, ef), eintölu (et) og fleirtölu (ft), án greinis? Svaraðu í *JSON formi eingöngu* og auðkenndu tölur og föll með skammstöfunum et, ft, nf, þf, þgf, ef."}], "ideal": "{\"et\": {\"nf\": \"hliðhollt lyfjapróf\", \"þf\": \"hliðhollt lyfjapróf\", \"þgf\": \"hliðhollu lyfjaprófi\", \"ef\": \"hliðholls lyfjaprófs\"}, \"ft\": {\"nf\": \"hliðholl lyfjapróf\", \"þf\": \"hliðholl lyfjapróf\", \"þgf\": \"hliðhollum lyfjaprófum\", \"ef\": \"hliðhollra lyfjaprófa\"}}"} {"input": [{"role": "system", "content": "Þú ert sérfræðingur í íslenskri málfræði."}, {"role": "user", "content": "Hvernig fallbeygist nafnliðurinn \"refsiverð stjörnuleit\" í öllum föllum (nf, þf, þgf, ef), eintölu (et) og fleirtölu (ft), án greinis? Svaraðu í *JSON formi eingöngu* og auðkenndu tölur og föll með skammstöfunum et, ft, nf, þf, þgf, ef."}], "ideal": "{\"et\": {\"nf\": \"refsiverð stjörnuleit\", \"þf\": \"refsiverða stjörnuleit\", \"þgf\": \"refsiverðri stjörnuleit\", \"ef\": \"refsiverðrar stjörnuleitar\"}, \"ft\": {\"nf\": \"refsiverðar stjörnuleitir\", \"þf\": \"refsiverðar stjörnuleitir\", \"þgf\": \"refsiverðum stjörnuleitum\", \"ef\": \"refsiverðra stjörnuleita\"}}"} {"input": [{"role": "system", "content": "Þú ert sérfræðingur í íslenskri málfræði."}, {"role": "user", "content": "Hvernig fallbeygist nafnliðurinn \"japönsk landbúnaðarvara\" í öllum föllum (nf, þf, þgf, ef), eintölu (et) og fleirtölu (ft), án greinis? Svaraðu í *JSON formi eingöngu* og auðkenndu tölur og föll með skammstöfunum et, ft, nf, þf, þgf, ef."}], "ideal": "{\"et\": {\"nf\": \"japönsk landbúnaðarvara\", \"þf\": \"japanska landbúnaðarvöru\", \"þgf\": \"japanskri landbúnaðarvöru\", \"ef\": \"japanskrar landbúnaðarvöru\"}, \"ft\": {\"nf\": \"japanskar landbúnaðarvörur\", \"þf\": \"japanskar landbúnaðarvörur\", \"þgf\": \"japönskum landbúnaðarvörum\", \"ef\": \"japanskra landbúnaðarvara\"}}"} {"input": [{"role": "system", "content": "Þú ert sérfræðingur í íslenskri málfræði."}, {"role": "user", "content": "Hvernig fallbeygist nafnliðurinn \"dýrmætt vistheimili\" í öllum föllum (nf, þf, þgf, ef), eintölu (et) og fleirtölu (ft), án greinis? Svaraðu í *JSON formi eingöngu* og auðkenndu tölur og föll með skammstöfunum et, ft, nf, þf, þgf, ef."}], "ideal": "{\"et\": {\"nf\": \"dýrmætt vistheimili\", \"þf\": \"dýrmætt vistheimili\", \"þgf\": \"dýrmætu vistheimili\", \"ef\": \"dýrmæts vistheimilis\"}, \"ft\": {\"nf\": \"dýrmæt vistheimili\", \"þf\": \"dýrmæt vistheimili\", \"þgf\": \"dýrmætum vistheimilum\", \"ef\": \"dýrmætra vistheimila\"}}"}