Source record

Published by arXiv cs.HC (T1 source). The original is at https://arxiv.org/abs/2608.21602.

Pipeline notes

The summary and note below are generated by the signal pipeline — they are Beyond Desk’s reading, not quotations from the source.

Preliminary empirical study evaluating how GPT-5, GPT-4o, GPT-4, and Claude Sonnet 4.6 detect and repair data issues that cause visualization defects, comparing single-agent versus multi-agent orchestration across zero-shot, guided identification, and guided repair prompting on a 911 emergency-call dataset with five injected issues.

Concrete evidence of where LLM agents succeed (single-field issues) and fail (temporal, geographic, semantic), with usable design implications for multi-agent visualization pipelines.

Cited by

No citations on record.