LegalBench-RAG: A Benchmark for Retrieval-Augmented Generation in the Legal Domain

Nicholas Pipitone, Ghita Houir Alami

2024arXiv preprint; peer-reviewed venue not established

What it adds

Transforms existing legal annotations into retrieval queries with character-span targets.

Table 1 totals 6,858; Table 2 totals 6,889. ContractNLI counts differ (946/977).

Selected statements located in the linked primary sources. These results have not been independently reproduced here.

The corpus draws from ContractNLI, CUAD, MAUD and PrivacyQA.
3.1 and Table 1
Experiments use the 776-query miniature benchmark, not the full corpus.
3.4; 4; Table 3

Cites

Converts clause annotations to retrieval targets.

Both evaluate retrieval for legal applications, using different retrieval units and relevance criteria.

Entry status, source edition & rights

Index entry: Source-linked summary, prepared with agent assistance. Independent scholarly review is pending.
Archive admission: This bibliographic entry does not confer admission to the versioned archive.
Edition inspected: 2408.10343v1
Reading coverage: Main sections 2–5 and Tables 1–7; tables 4–5 inspected as rendered pixels; author repository README and file tree.
Rights: External references only. Papers and research artifacts remain with their original hosts; their own terms apply.
Entry updated: 2026-10-15

Source-linked summary · independent assessment pending.