seedeval

Building a benchmark to evaluate the diversity of seeds generated using different tool calls.

Overview

• Environment ID: seedeval
• Short description:
• Tags: <tool-use, synthetic-data>
• Source: charvibannur/seed-eval-benchmark

Datasets

• Primary dataset(s): <Fine Web (Works with any text dataset)>

Architecture

seed-eval-benchmark/
└── environments/
    └── seedeval/
        ├── README.md
        ├── pyproject.toml
        └── seedeval.py

Quickstart

Run an evaluation with default settings:

prime eval run seedeval

Configure model and sampling:

prime eval run seedeval -m gpt-4.1-mini  -n 20 -r 3

Metrics

Summarize key metrics your rubric emits and how they’re interpreted.