3 ms·
Anthropic’s tool calling was exposed as XML tags at the beginning, before they introduced the JSON API. I expect they’re still templating those tool calls into
by wolttam 8mo ago
Anthropic’s tool calling was exposed as XML tags at the beginning, before they introduced the JSON API. I expect they’re still templating those tool calls into XML before passing to the model’s context
- pocketarc 8mo agoYeah like I remember prior to reasoning models, their guidance was to use <think> tags to give models space for reasoning prior to an answer (incidentally, also the reason I didn't quite understand the fuss with reasoning models at first). It's always been XML with Anthropic.
- wolttam 8mo agoExactly the same story here. I still use a tool that just asks them to use <think> instead of enabling native reasoning support, which has worked well back to Sonnet 3.0 (their first model with 'native' reasoning support was Sonnet 3.7)
- scotty79 8mo agoCan you sniff it out with Wireshark?
- wolttam 8mo agoThey don't expose the raw context over the wire, it's all pre/post processed at their API endpoints.