Glossary term
Glossary term
Agentic Systems
A prompt having a single correct response. For example, consider the following prompt:
True or false: Saturn is bigger than Mars.
The only correct response is true.
Contrast with no one right answer.
Created for this library
A factual QA system labels its evaluation set as one-right-answer because each question has a single correct answer the model must produce.
An LLM evaluation team treats math problems as one-right-answer prompts and grades against the exact final value.
A code-generation team uses one-right-answer evaluation on coding benchmarks where the correctness criterion is whether tests pass.
Definition source: Google for Developers Machine Learning Glossary | Creative Commons Attribution 4.0 License