Search papers, labs, and topics across Lattice.
2
0
5
3
Slash LLM evaluation costs by over 270x without sacrificing accuracy, revealing that even top open-weight models struggle to outperform simple baselines on AML transaction fraud detection.
Prompt engineering's dead end? This study finds that model capability trumps prompt-level interventions for mathematical reasoning, suggesting that focusing on model selection and verification may be more fruitful than further prompt optimization.
