Hacker GitHub

← All repos

gkamradt/needle-in-a-haystack

View on GitHub ↗ ★ 2,407 Jupyter Notebook

Doing simple retrieval from LLM models at various context lengths to measure accuracy

4 mentions on Hacker News by 4 people: 1 story, 3 comments · first 2023-11-21 · latest 2026-02-01

Mentions per year 4 mentions · busiest year 2024 (2)
20082026
AllStoriesComments NewestTopOldest
  1. …[4] https://github.com/nicobailon/pi-mcp-adapter [5] https://github.com/steipete/mcporter [6] https://github.com/gkamradt/LLMTest_NeedleInAHaystack [7] https://research.trychroma.com/context-rot [8] https://arxiv.org/html/2601.20834v1
    comment by mustaphah · 2026-02-01 · in What I learned building an opinionated and minimal coding agent
  2. GPT-4 has much worse recall compared to Claude-3 though, compare these two haystack tests: https://github.com/gkamradt/LLMTest_NeedleInAHaystack/raw/ma... https://cdn.arstechnica.net/wp-content/uploads/2024/03/claud...
    comment by danielbln · 2024-03-27 · in Claude 3 beats GPT-4 on Aider's code editing benchmark
  3. Yep that's pretty much it! That's what they call needle in a haystack. See: https://github.com/gkamradt/LLMTest_NeedleInAHaystack
    comment by halflings · 2024-02-15 · in Our next-generation model: Gemini 1.5
  4. ▲ 1
    Analysis to test in-context retrieval ability of GPT-4-128K context
    story by kvee · 2023-11-21 · 0 comments