Seed Run Comparator
Compare repeated ML runs using matched seeds, clear dispersion, and a check for missing or failed runs. Produce a reproducible results table.
$1
Works with the AI Agents you already use
Seed Run Comparator
Example session with this skill installed
Compare baseline seed scores 1:70, 2:71, 3:69 against candidate 1:73, 3:71, 4:75.
- Read your context and instructions
- Compiled the seed run comparator
Illustrative output
Seeds 1 and 3 are paired. Their candidate-minus-baseline deltas are +3 and +2, averaging +2.5. Seed 2 is missing for the candidate and seed 4 for the baseline. Report 2 matched pairs and retain full-group summaries separately. Two pairs are insufficient for a broad reliability claim.
Connects securely to your tools. The creator never sees your data.
About this skill
Identify the evaluation metric, direction, task, dataset split, config, seed and run status. Verify the config and data match except the named treatment. Deduplicate by run ID and flag multiple completed runs for the same config and seed; apply a provided selection rule or surface the conflict. Exclude only according to an explicit reason and retain failed or missing runs in coverage.
Compute per-config mean, sample standard deviation when n exceeds 1, and n. When seeds are intended to match, report the intersection and paired candidate-minus-baseline deltas. Present an unpaired comparison when matching is unavailable. Avoid pooling heterogeneous tasks unless aggregation weights are defined. Use percentage points for raw changes on percentages.
Return a compact table, paired differences, exclusions, and a conclusion scoped to the observed sample. Provide exact inputs or formulas needed to reproduce the reported values. Mark dispersion as unavailable for a single run and avoid inferring statistical significance from overlapping standard deviations.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed today
Permissions
No special permissions declared or detected