PhantomBench
Explore score history ↗PhantomBench evaluates language models' ability to abstain from answering about non-existent entities. It comprises over 60,000 non-existent terms derived from real concepts across domains, and provides a pipeline for generating further instances.
- Released
- 2026-06-09
- Institutions
- University of British Columbia · Canadian Institute for Advanced Research
- Evaluates
- General AI, Language & Knowledge, cs.CL
- Openness
- unknown
- Importer
- Claire Radar
- Review status
- unreviewed
PhantomBench paper
No reported scores are on record for this benchmark yet.