I fine-tuned a 1.5B model into a groundedness judge on my laptop

I fine-tuned a 1.5B model into a groundedness judge on my laptop — and itI fine-tuned a 1.5B model into a groundedness judge on my laptop — and it

The network for creativity

Join 1.25M professional creatives like you

Connect with clients, get discovered, and run your business 100% commission-free

Creatives on Contra have earned over $150M and we are just getting started

Back to feedPost

Daniel Enache

• Jul 4

I fine-tuned a 1.5B model into a groundedness judge on my laptop — and it matched a frontier judge 100% of the time, for $0 per call.

Every RAG or agent eval gate needs a judge: something that reads a (question, context, answer) triple and rules whether every claim is actually supported. Usually that judge is a frontier-model API call — so every CI run costs money, leaks data off-box, and gets rate-limited.

So I distilled that capability into a model that runs locally, in CI, on every commit.

groundcheck (open source): • Dataset with labels correct by construction — no human labelling, no LLM labelling. • QLoRA fine-tune of a 1.5B model, 4-bit, ~4.5 min on an 8 GB laptop GPU. • The rule: the fine-tune ships only if the eval suite says it beats the baseline.

Results on held-out cases the model never saw: • base model, zero-shot: F1 0.083 (it hedges; handles refusals 0% of the time) • fine-tuned: F1 1.000

Then the question that actually matters — does it match a frontier judge? On 60 held-out cases, my local judge and a frontier teacher both scored 1.000 accuracy, with 100% agreement. Cost: $0.026 / 1,000 calls for the frontier judge, $0.00 for mine.

Honest scope: the corpus is synthetic (labels correct by construction) — this proves the pipeline end to end. Next: real RAG traces.

The loss curve doesn't decide what ships. The eval suite does. Repo + results → https://github.com/enached134-ctrl/groundcheck

github.com

GitHub - enached134-ctrl/groundcheck

Contribute to enached134-ctrl/groundcheck development by creating an account on GitHub.

Razvan Vezeteu

pro

• 2d

SaaS gradient blobs don't fail because they look bad. They fail because nobody remembers them. Safe disappears, on purpose or not. Went with texture and contrast instead.

Open to projects if this is more your speed.

Illustration noise lookingforwork

Zhan Y.

• 2d

These are really cool, Rezvan!

Jacob Weinstein

• 2d

A chart for my hand built quantitative model’s performance in the last 30 days, as well as live alerts of any changes the model picked up in its last data scan.

AI Application Development Web Scraping Web Development Claude Node.js TypeScript

Olatunji Yusuf

• 2d

Developed and tested a proprietary trading automation system over a two-week period. The system focuses on disciplined risk management, automated execution, and consistency. During the test period, the account grew from approximately $2,000 to $8,000 while following predefined trading rules and market conditions. Results shown are from a historical test period and are not a guarantee of future performance.