SPINE Benchmark
Explore score history ↗SPINE evaluates LLM sycophancy by having an adaptive LLM proxy persistently challenge a target model across up to 25 turns on false-presupposition and unethical-query items, measuring collapse rates.
- Released
- 2026-09-08
- Institutions
- Texas A&M University · University of Cincinnati
- Evaluates
- General AI, Language & Knowledge, cs.CL
- Openness
- unknown
- Importer
- Claire Radar
- Review status
- ai-reviewed
SPINE paper
No reported scores are on record for this benchmark yet.