The current LLM prompts were developed for non-reasoning models. Reasoning models should give better performance for this task, but may require different prompts. (E.g. no chain-of-thought prompting).
Current thinking is to target GPT o3 (proprietary) and Nemotron 3 Nano (open source).
The current LLM prompts were developed for non-reasoning models. Reasoning models should give better performance for this task, but may require different prompts. (E.g. no chain-of-thought prompting).
Current thinking is to target GPT o3 (proprietary) and Nemotron 3 Nano (open source).