I'm doing coreference resolution and this model (w/o thinking) performs at the G... | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

		sync 18 days ago \| parent \| context \| favorite \| on: DeepSeek-v3.1 I'm doing coreference resolution and this model (w/o thinking) performs at the Gemini 2.5-Pro level (w/ thinking_budget set to -1) at a fraction of the cost.

antman 18 days ago | [–]

Nice point. How did you test for coreference resolution? Specific prompt or dataset?

dr_dshiv 18 days ago | [–]

Strong claim there!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact