entity · derived
Large language models
Derived node: assembled mechanically from the claims carrying large-language-models. A roster, not an adjudicated definition.
Every claim under this term
- 5541658-032 : Benchmark results for legal LLMs may overstate capability because of data contamination: if a model saw a benchmark's ground truth answers during training, its measured performance reflects memorizati
- 5541658-034 : Even as LLMs for judicial reasoning become more sophisticated, their inability to replicate empathy and social responsiveness will confine them to supportive rather than decisional functions.