Model Card Summary Builder

How Model Card Summary Builder works

Press Load model once: your browser downloads Gemma 4 E2B instruction-tuned (web build) (2.0 GB) from this site and keeps it in its cache. Then add your text and press Preview local signals. The model runs on this device using your graphics card, so nothing you enter is sent to a server. Until the model is loaded, or if your device cannot run it, the tool falls back to a simple keyword check and labels the result as the fallback.

Gemma 4 E2B instruction-tuned (web build) is a small open language model from Google. The tool sends it the instruction below plus your text, asks for a JSON answer with fixed fields, and shows each field as a card. It runs at temperature 0, so the same input gives the same answer on the same device, though different graphics cards can word it slightly differently.

Inputs longer than 6,000 characters are cut to that length before they reach the model, and the result says so. The model only sees what you paste: it does not look anything up.

The instruction the model gets

You review draft model cards. Summarise what the text says about the model and point out which standard model-card sections are missing or thin: intended use, training data, evaluation data and metrics, known limitations, bias and fairness, and out-of-scope uses. Quote numbers exactly as written; never invent metrics, datasets or results.

What it returns

• Summary: Two or three sentences on what the model is, what it is for and how it was evaluated, as the text states.
• Evaluation as stated: Evaluation data and results exactly as the text gives them.
• Limitations mentioned: Limitations, risks or known failure cases the text mentions.
• Missing or thin sections: Model-card sections the text does not cover or covers vaguely (intended use, training data, evaluation, limitations, bias and fairness, out-of-scope uses).
• Questions for the model owner: Specific questions a reviewer should ask before publishing this card.

Worked example

The model's output for the built-in sample, recorded on a desktop browser. Your run may word things slightly differently.

Model card draft: the classifier was evaluated on internal support tickets and public benchmark samples. Accuracy improved after filtering low-quality labels, but edge-case performance for multilingual inputs still needs human review before launch.

Summary

The classifier was evaluated on internal support tickets and public benchmark samples. Accuracy improved after filtering low-quality labels, but edge-case performance for multilingual inputs still requires human review before launch.

Evaluation as stated

evaluated on internal support tickets evaluated on public benchmark samples

Limitations mentioned

edge-case performance for multilingual inputs still needs human review

Missing or thin sections

intended use training data evaluation data and metrics known limitations bias and fairness out-of-scope uses

Questions for the model owner

What is the specific intended use of this classifier? What is the composition and quality of the training data? What specific metrics were used to measure the improvement in accuracy? What are the specific edge cases that require human review for multilingual inputs?

The model

The files are served from this site, versioned so a cached copy is never stale. Nothing is fetched from a third party while you use the tool.

Limitations

• Small language models make mistakes: they can miss an item, merge two, or state something more confidently than the text supports. Check the result against your input before you rely on it.
• It needs WebGPU and about 8 GB of memory, so it is offered on desktops and laptops, not on phones and tablets. Devices that cannot run it get the keyword fallback, clearly labelled.
• Results describe the input you gave, not the real-world situation behind it. Review them yourself before acting.

Good uses

• Triage model-card drafts before deeper review
• Prepare consistent review labels for repeat workflows
• Surface missing evaluation and caveat signals
• Keep sensitive release notes in the browser preview