News for hackers. Stuff that matters.
Re:A prompt-trained DeepSeek R1 70B can perform better than ... by meame2010 on Wednesday February 05, 2025 @01:38PM (#42953085)
Time to move to open-source and smaller reasoning model.

Here are the top three learnings from auto-prompt optimizing DeepSeek R1 LLaMA70B for RAG:

1⃣ A trained DeepSeek R1 LLaMA70B(r1 distilled) is even better than GPT-o1 without training. 2⃣ The “Reasoning” model is less susceptible to overfitting compared with non-reasoning models. By comparing it with GPT-3.5, both gpt3.5 and r1 distilled start at the same accuracy and reach similar accuracy on the validation dataset. However, on the test dataset, r1 distilled often achieves much higher accuracy. 3⃣ R1 can think too long and run out of output tokens before finishing the task. The optimized prompt specifically added instructions for it to “think less.”

Re:A prompt-trained DeepSeek R1 70B can perform better than ... by ai-christianson on Wednesday February 05, 2025 @02:28PM (#42953788)
So this is mainly optimizing the prompt itself? E.g. RAG + optimizing few shot examples?
Re:A prompt-trained DeepSeek R1 70B can perform better than ... by meame2010 on Wednesday February 05, 2025 @03:05PM (#42954340)
not few shot, but prompt tuning via text generation via auto-differentiation.

https://arxiv.org/abs/2501.16673

Re:A prompt-trained DeepSeek R1 70B can perform better than ... by grapeghost on Wednesday February 05, 2025 @01:55PM (#42953339)
This is really exciting news! I can't wait to see how things progress from here. I'd love to see some cost comparisons alongside the performance gains in the future.