HalluLens

A hallucination benchmark (Bang et al., ACL 2025) that classifies model hallucination into intrinsic and extrinsic types. Met by this vault via claim-naser-2026-citation-fabrication-audit-no-frankfurt-bullshitting-citation, which cites it as reference [18] alongside TruthfulQA among the two hallucination benchmarks named in Naser (2026)'s reference list; under HalluLens's taxonomy, citation fabrication would classify as extrinsic hallucination, per Naser's citation, though the vault has not yet read HalluLens directly or checked its own citation of the Frankfurt/bullshitting literature.

Logged as a watching stub rather than a full hub: it has appeared in exactly one paper's reference list so far, encountered secondhand rather than read. Worth revisiting if a second source in the vault cites it directly or if it is read at the primary.

References

• claim-naser-2026-citation-fabrication-audit-no-frankfurt-bullshitting-citation
• Capture: 10-inbox/raw/2026-11-04-does-any-technical-ml-paper-on-citationreference-fabrication.md