Evaluation Bench
Practical run logs, evaluation methods, build reports, and postmortems from LoRADock contributors.
Evaluation Bench
5 articlesComparing Two Ticket-Triage Adapters on the Same 500 Requests
A practical LoRADock community note on inputs, validation, tradeoffs, and release decisions for comparing two ticket-triage adapters on the same 500 requests.
Why Training Loss Did Not Predict Production Quality
A practical LoRADock community note on inputs, validation, tradeoffs, and release decisions for why training loss did not predict production quality.
A Small Regression Suite for Text-to-Image LoRAs
A practical LoRADock community note on inputs, validation, tradeoffs, and release decisions for a small regression suite for text-to-image loras.
Measuring Structured JSON Reliability Without an LLM Judge
A practical LoRADock community note on inputs, validation, tradeoffs, and release decisions for measuring structured json reliability without an llm judge.
Human Review Notes from 1,000 Generated Product Descriptions
A practical LoRADock community note on inputs, validation, tradeoffs, and release decisions for human review notes from 1,000 generated product descriptions.