muhammad-junaid · portfolio

Small Language Model Spec Experiments

Where does a small, local coding model stop producing numerically correct business logic?

An empirical study of a small, locally-hosted coding model (Qwen2.5-Coder-3B on Apple MLX, 8 GB M1) and the point where it stops generating numerically correct business logic.

A five-rung spec ladder holds the rules and expected outputs constant and varies only how the spec is phrased, which pins the exact failure boundary: the model handles plain prose fine but breaks the moment literal lookup tables are removed, not on the multi-tier and boolean constructs presumed to be the culprits.

Re-phrasing the same ten rules in the model-friendly dialect moved it from 1 of 12 correct to 12 of 12 on the first attempt, and the result holds at three times the rule count (23 of 23).

  • Python
  • Apple MLX
  • Qwen2.5-Coder
  • LoRA fine-tuning
  • README.md
  • how-i-work.md
  • fix-me.ts
  • wheelbuilds.tsxengineering
  • inspect.tsxengineering
  • identimal.tsxengineering
  • local-coding-model-exp.tsxengineering
  • this-portfolio.tsxengineering
  • crowdfunding-platform.mdproduct
  • web3-nft-dapp.previewproduct
  • real-estate-platform.previewproduct
  • enterprise-risk-management.mdproduct
  • tech-recruiting-platform.mdproduct
  • salesforce-lightning-builds.mdproduct
  • b2b-procurement-platform.mdproduct
  • enterprise-portals.mdproduct
  • shipping-with-ai.mdproduct
  • skills.json
  • contact.ts
  • Muhammad_Junaid_Resume.pdf
  • Switch theme
  • Toggle terminal
  • Open assistant
  • Download resume
  • Copy emailmjunaid7843@gmail.com