feat(eval): Implement meta-evaluation #66
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
| name: Tests | |
| on: | |
| pull_request: | |
| paths: | |
| - .github/workflows/tests.yml | |
| - safeguards/** | |
| - pyproject.toml | |
| - tests/** | |
| jobs: | |
| lint: | |
| name: check linting using ruff | |
| runs-on: ubuntu-latest | |
| steps: | |
| - uses: actions/checkout@v4 | |
| - uses: chartboost/ruff-action@v1 | |
| # test-secrets-guardrail: | |
| # name: test secrets guardrail | |
| # runs-on: ubuntu-latest | |
| # steps: | |
| # - uses: actions/checkout@v4 | |
| # - run: | | |
| # pip install -U pip | |
| # pip install torch --index-url https://download.pytorch.org/whl/cpu | |
| # pip install -e ".[dev]" | |
| # python -m pytest -s tests/test_secrets_detection.py | |
| test-privilege-escalation: | |
| name: test privilege escalation guardrail | |
| runs-on: ubuntu-latest | |
| env: | |
| OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }} | |
| steps: | |
| - uses: actions/checkout@v4 | |
| - run: | | |
| pip install -U pip | |
| pip install torch --index-url https://download.pytorch.org/whl/cpu | |
| pip install -e ".[dev]" | |
| python -m pytest -s tests/test_privilege_escalation.py | |
| test-prompt-injection-llm: | |
| name: test prompt injection llm guardrail | |
| runs-on: ubuntu-latest | |
| env: | |
| OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }} | |
| steps: | |
| - uses: actions/checkout@v4 | |
| - run: | | |
| pip install -U pip | |
| pip install torch --index-url https://download.pytorch.org/whl/cpu | |
| pip install -e ".[dev]" | |
| python -m pytest -s tests/test_prompt_injection_llm.py | |
| test-prompt-injection-classifier: | |
| name: test prompt injection classifier guardrail | |
| runs-on: ubuntu-latest | |
| env: | |
| HF_TOKEN: ${{ secrets.HF_TOKEN }} | |
| WANDB_API_KEY: ${{ secrets.WANDB_API_KEY }} | |
| steps: | |
| - uses: actions/checkout@v4 | |
| - run: | | |
| pip install -U pip | |
| pip install torch --index-url https://download.pytorch.org/whl/cpu | |
| pip install -e ".[dev]" | |
| huggingface-cli login --token ${{ secrets.HF_TOKEN }} | |
| python -m pytest -s tests/test_prompt_injection_classifier.py | |
| test-evaluation-classification: | |
| name: test evaluation classification | |
| runs-on: ubuntu-latest | |
| env: | |
| WANDB_API_KEY: ${{ secrets.WANDB_API_KEY }} | |
| steps: | |
| - uses: actions/checkout@v4 | |
| - run: | | |
| pip install -U pip | |
| pip install torch --index-url https://download.pytorch.org/whl/cpu | |
| pip install -e ".[dev]" | |
| python -m pytest -s tests/test_evaluation_classification.py |