---
pretty_name: Fara 1.5 9B CUA Baseline Evaluation Artifacts
configs:
- config_name: task-results
  data_files: data/task-results-*.parquet
  default: true
- config_name: runs
  data_files: data/runs-*.parquet
---

# Fara 1.5 9B baseline evaluation artifacts

This public dataset preserves the complete finalized artifacts for the `fara15-dom-baseline` run of WootzApp CUA Bench. It was published under the `WootzappLab` organization by Hugging Face account `ishagarg1103` (IshaGarg1103).

## Result

| Metric | Value |
|---|---:|
| Tasks selected | 90 |
| Passed | 6 |
| Verifier failures | 60 |
| Execution errors | 24 |
| Infrastructure errors | 0 |
| Pass@1 over all selected tasks | 6.67% |
| Pass rate among verified tasks | 9.09% |

Execution errors remain in the denominator. The raw logs are included so those 24 tasks can be audited separately from verifier-scored failures.

## Exact provenance

| Component | Repository or resource | Exact revision |
|---|---|---|
| Upstream model | [`microsoft/Fara1.5-9B`](https://huggingface.co/microsoft/Fara1.5-9B) | `1a93677cd89d5601bc2ed759791e981f3a520032` |
| Preserved model checkpoint | [`WootzappLab/fara-9b-baseline-eval`](https://huggingface.co/WootzappLab/fara-9b-baseline-eval) | `6b15345b4dd7d567863bc08ec3bc79b44780dd99` |
| Fara inference source | Microsoft Fara source used by the serving setup | `a675d6d61c41c47ae87bacefeab22caad18e3e84` |
| Task dataset | [`WootzappLab/cua-bench`](https://huggingface.co/datasets/WootzappLab/cua-bench) | `745eebb8f07abb92fdaae356bb6bee3a1f2975af` |
| Recorder, browser harness, environments, verifier | [`tokenbender/w8-cua-bench` `environment-work`](https://github.com/tokenbender/w8-cua-bench/tree/environment-work) | `66037db6bae4b00a2b5860aa08ebb69b559e6871` |
| Infrastructure code | Infra branch/PR | Pending publication; no SHA is claimed |
| Evaluation content identity | Core result bundle SHA256 | `3469791e6d4443748b8e0e4438a0770b6c05c927725b777c55bab574b50df880` |

The evaluation ran on GCP instance `w8-fara-9b-8115f370-head-3ibtw2ah-compute` in project `lifeandhalf-24122025`, zone `us-central1-a`.

## Repository layout

- `artifacts/3469791e6d4443748b8e0e4438a0770b6c05c927725b777c55bab574b50df880/evaluation-runs/fara15-dom-baseline/`: complete raw run output, including per-task recordings, model output, logs, rubrics, and verifier results
- `data/task-results-00000-of-00001.parquet`: one normalized row per task
- `data/runs-00000-of-00001.parquet`: run-level summary and provenance
- `metadata/evaluation.json`: machine-readable evaluation receipt
- `metadata/runs/fara15-dom-baseline.json`: run receipt with all task results
- `metadata/publication.json`: publication identity
- `manifests/3469791e6d4443748b8e0e4438a0770b6c05c927725b777c55bab574b50df880.sha256`: SHA256 for every raw artifact file

## Integrity check

After downloading the repository, run:

```bash
sha256sum -c manifests/3469791e6d4443748b8e0e4438a0770b6c05c927725b777c55bab574b50df880.sha256
```

The content identity is derived from `run_config.json`, `results.jsonl`, and `summary.json`. It is a SHA256 digest, not a Git commit. The immutable Hugging Face repository commit is recorded by the `eval-fara15-dom-baseline` tag.

## Scope

This repository contains evaluation evidence and structured results. It does not contain access tokens, TLS certificates, browser caches, or model weights. The weights are preserved separately in the checkpoint repository linked above.