# Evaluation Results Template

Use this file after loading the base model or a trained adapter locally. Do not fill it with expected results; record only observed results.

## Run Metadata

```text
Date:
Evaluator:
Base model:
Adapter path:
Commit/hash:
Hardware:
Runtime:
Precision:
Command:
```

## Package Checks

```text
scripts/verify_package.py: PASS/FAIL
scripts/check_local_readiness.py: PASS/FAIL
scripts/verify_hf_license.py: PASS/FAIL/NOT RUN
```

## Local Load

```text
Base model load: PASS/FAIL
Adapter load: PASS/FAIL/NOT APPLICABLE
Smoke prompt result: PASS/FAIL
```

## Benign Freedom Eval

```text
Command:
answered:
possible_refusal:
empty:
Notes:
```

## Quality Notes

```text
Spanish:
Coding:
Defensive security:
Legal/medical caveats:
Unexpected refusals:
Unexpected unsafe outputs:
```

## Publication Decision

```text
Ready to publish: yes/no
Reason:
Remaining blockers:
```