Documents created by those much more experienced than I. This directory is meant to give beginner builders context of the limitations of an AI's "thought" process.
Highlights:
- Lost In the Middle: How Language Models Use Long Contexts
- Abstract: While recent language models have the ability to take long contexts as input, relatively little is known about how well they use longer context. We analyze the performance of language models on two tasks that require identifying relevant information in their input contexts: multi-document question answering and key-value retrieval.
- "Needle In A Haystack" Analysis: Pressure Testing GPT-4 & Claude 2.1's Long Context Retrieval Accuracy
- An exported email from Greg Kramradt's mailing list. He personally pressure tested these models with the intention of helping others make smarter decisions when it comes to retrieval.
A directory to hold file-base files.
Currently only contains the Assistant Architect's files which may be useful for inspiration or insight.
The files available may be turned into helpful templates, or if you're unsatisfied with how AA4LLM works, you might learn what not to do for your own Custom GPT.