gskm

3 ms·

1. 1.▲
   by gskm 7mo ago

   Thanks! Glad it's working well for you. A few practical tips:
2. 2.▲
   by gskm 7mo ago

   Update on the benchmark numbers: the results in the original post were computed with a looser tokenizer, making the budget less strict than it should be. We've since improved that — the budget is now accurate end-to-end. Corrected numb
3. 3.▲
   by gskm 7mo ago

   Thank you! That's exactly the goal, drop-in token savings without changing your LLM pipeline. If you give it a spin, I'd love to hear how it works on your data. We're actively tuning the ranking based on early feedback, so an
4. 4.▲
   Show HN: HighSNR – Cut length and noise from your LLM context(high-snr.com)

   6 points by gskm 7mo ago|5 comments
5. 5.▲
   by gskm 7mo ago

   Loved this article. I'd add a few things I wish someone had told me when I was starting my PhD: 1) Maximize variance, but know when to stop. Karpathy's point is great. Explore early, say yes to different things. But at some point
6. 6.▲
   by gskm 7mo ago

   I think your instinct is right. More context isn't free, even when the window supports it, and the model still has to attend to everything in there, and noise dilutes the signal. A cleaner, smaller context consistently gives better out
7. 7.▲
   by gskm 7mo ago

   The context-overload rule resonates — we kept hitting the same problem. Diagnosis is useful but we ended up just compressing the retrieved chunks to a token budget before they hit the LLM. Deterministic, keeps only the highest-signal passag