stansApprentice

3 ms·

1. 1.▲
   TrajectoryKit, a competitive open-source deep research agent without fine-tuning(williamlugoloobi.com)

   1 points by stansApprentice 7mo ago|0 comments
2. 2.▲
   by stansApprentice 8mo ago

   Thank you! I haven't used Komilion before, but it seems like we were thinking in the same direction. There's definitely room to add more to our suggested application of the probe. That's why we kept it simple so that people c
3. 3.▲
   by stansApprentice 9mo ago

   Author here. The core idea is pretty simple: train linear probes on the model's internal state before it generates anything to predict if it'll succeed. Then use those predictions to route queries: Send easy ones to cheap infere
4. 4.▲
   Routing LLM queries using internal success predictions (70% cost reduction)(arxiv.org)

   1 points by stansApprentice 9mo ago|3 comments
5. 5.▲
   LLMs encode how difficult problems are(arxiv.org)

   174 points by stansApprentice 1y ago|38 comments