2 ms·
Apparently you can do the same by simply running it without reasoning, while giving it a thinking tool... >guys you do know you can just disable thinking, and
by Pragmata 2mo ago
Apparently you can do the same by simply running it without reasoning, while giving it a thinking tool...
>guys you do know you can just disable thinking, and instead give it a "deep_think" tool, and it will call it with internal CoT reasoning format right?
>gl fixing that
https://x.com/_can1357/status/2087228354399265125?s=20 https://x.com/_can1357/status/2087228354399265125?s=20
- retinaros 2mo agoits not exactly the same... its tool use spec asking to put thinking in inputs fields... it is a good idea but its not same.
- ashirviskas 2mo agoI've been doing that since before reasoning was a thing baked into the models, it always performs better this way. Except for some providers/models where you just can't easily turn it off, now I just avoid them. This way I save tokens and have full control of the reasoning.
- MaxMatti 2mo agoHow does it save tokens?
- klntsky 2mo agoThe model is finetuned to enter/leave its thinking mode using special token separators. there's no reason to assume the tool calls induce the same token distribution or produce the model's actual native reasoning trace
- its-summertime 2mo agoWe have its actual reasoning traces, and we have these psudotraces, distribution / nativeness is testable now