VLP adds an NL documentation layer with trace-linked mismatch detection and derived formal checks to make human validation of LLM code feasible, lifting pass@1 from 28.7-73.2% to 65.4-93.5%.
Large language models should ask clarifying questions to increase confidence in generated code,
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
verdicts
UNVERDICTED 2representative citing papers
Introduces NoisyToolBench benchmark and Ask-when-Needed framework to improve LLM tool-use performance when user instructions are unclear or incomplete.
citing papers explorer
-
Guiding Human Validation of LLM-Generated Code via Verifiable Literate Programming
VLP adds an NL documentation layer with trace-linked mismatch detection and derived formal checks to make human validation of LLM code feasible, lifting pass@1 from 28.7-73.2% to 65.4-93.5%.
-
Learning to Ask: When LLM Agents Meet Unclear Instruction
Introduces NoisyToolBench benchmark and Ask-when-Needed framework to improve LLM tool-use performance when user instructions are unclear or incomplete.