quiz bu lëkkale ak tegtal · Diggante Niveau
Version bu gaaw ak saytu régression
Dafay saytu li bokk ci taabal bu gaaw buñ joxe, naka la suite regression di doxalee ci nondeterminism, ak lu tax model snapshots ak àttekati LLM soxla pinning.
Laaj 1 de 8
Ban mbir lañu wara boole muy benn configuration prompt?
Xoolaatal laaj bu nekk8 laaj
Sooy def quiz bi, tontu yi ak leeral yi dañuy feeñ.
- Ban mbir lañu wara boole muy benn configuration prompt?
- Only the prompt wording
- Tur model bi kese
- Prompt template, model identifier, sampling parameters and tool schemas
- Only the user inputs from production
- Lan moo tax nga takk nataalu model bu am bis ci barabu pseudo-alias bu mujj?
- So model changes happen only when you deliberately switch and re-test
- Because aliases are always slower
- Because snapshots are free
- Because aliases do not support system prompts
- Lu guide bi wax ci tàngoor 0?
- It guarantees identical outputs every time
- It makes regression testing unnecessary
- It only affects the length of outputs
- It reduces randomness but does not guarantee identical outputs across runs
- naka la suite régression wara def ak génne yi amul benn dogal?
- Require exact string matches
- Use thresholds and repeated samples instead of exact equality
- Ignore any failing cases
- Only test with a single example
- Lan mooy gëna am solo ci test cases yu bees ngir set wurus?
- Past production failures turned into test cases
- Random text from the web
- Only cases the current prompt already passes
- Outputs copied from a competitor
- If an LLM judge grades outputs, what should you do?
- Let the judge model change freely
- Trust every verdict without review
- Pin the judge's model and prompt and check it against human labels
- Use the same prompt under test as the judge
- Why log the prompt version ID with each production request?
- To reduce token costs
- To improve model accuracy
- To satisfy the model provider's terms
- To tie problems to a specific deploy and roll back quickly
- Why list individual cases that flipped from pass to fail rather than only the average score?
- Averages are always wrong
- An unchanged average can hide new breakage offset by gains elsewhere
- Flipped cases reduce API costs
- Only flipped cases are graded