Consensus And Disagreement Detection Across Literature

via “cross-paper-finding-synthesis-and-consensus-detection”

AI agent for automated systematic literature reviews.

Unique: Uses embedding-based clustering of extracted claims to identify consensus and disagreement patterns, then conditions LLM summaries on cluster statistics, rather than naively aggregating paper abstracts or using citation co-occurrence

vs others: More precise than citation network analysis because it operates on semantic claim content rather than citation patterns, and more scalable than manual meta-analysis because it automates finding extraction and clustering