An LLM-based cleaning and re-chunking pipeline for the BRIGHT benchmark improves retrieval scores, but the gain is dominated by chunking rather than noise removal.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
BRIGHT+: Upgrading the BRIGHT Benchmark with MARCUS, a Multi-Agent RAG Clean-Up Suite
An LLM-based cleaning and re-chunking pipeline for the BRIGHT benchmark improves retrieval scores, but the gain is dominated by chunking rather than noise removal.