Zeno 🤝 OpenAI Evals
Use Zeno to visualize the results of OpenAI Evals.
https://user-images.githubusercontent.com/4563691/225655166-9fd82784-cf35-47c1-8306-96178cdad7c1.mov
Example using zeno-evals to explore the results of an OpenAI eval on multiple choice medicine questions (MedMCQA)
Usage
pip install zeno-evals
Run an evaluation following the evals instructions. This will produce a cache file in /tmp/evallogs/.
Pass this file to the zeno-evals command:
zeno-evals /tmp/evallogs/my_eval_cache.jsonl
Example
Single example looking at US tort law questions:
zeno-evals ./examples/example.jsonl
And an example of comparison between two models:
zeno-evals ./examples/crossword-turbo.jsonl --second-results-file ./examples/crossword-turbo-0301.jsonl
And lastly, we can pass additional Zeno functions to provide more context to the results:
pip install wordfreq
zeno-evals ./examples/crossword-turbo.jsonl --second-results-file ./examples/crossword-turbo-0301.jsonl --functions_file ./examples/crossword_fns.py
Metadata
Release files for zeno-evals 0.1.10
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| zeno_evals-0.1.10.tar.gz | 4.7 kB | Without provenance | Details for zeno_evals-0.1.10.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| zeno_evals-0.1.10-py3-none-any.whl | Python 3 | none | any | Details for zeno_evals-0.1.10-py3-none-any.whl |
Total release size: 10.0 kB
Release files / zeno_evals-0.1.10.tar.gz
| Download URL | zeno_evals-0.1.10.tar.gz |
|---|---|
| Size | 4.7 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
3ecf12896512ed8ec8bff10cc628c0a840c74b3a7c90c8f068309cf79ac8dbc8
|
|
BLAKE2b-256 checksum How to use checksums |
f03ffd3e7dfa364afea27e0add96cfb03daeb9cee94f0632a80fffb2e5dcfe95
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
poetry/1.3.2 CPython/3.8.12 Darwin/22.4.0
|
Release files / zeno_evals-0.1.10-py3-none-any.whl
| Download URL | zeno_evals-0.1.10-py3-none-any.whl |
|---|---|
| Size | 5.3 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
9ee8451d68518fbeba4066f4c48fff2aef5e653299296d67521aa84a03cd677b
|
|
BLAKE2b-256 checksum How to use checksums |
a0f25688d4f1fd25eeaeb850511ac7b01b9fc6c0e9af044648922c56b59fa5d1
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
poetry/1.3.2 CPython/3.8.12 Darwin/22.4.0
|
Security for zeno-evals 0.1.10
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.