Curating Always-Loaded Context for LLM Agents: A Capacitated Assortment Model with Censored Feedback

Key points

• Each loaded token in the file is charged again in every later round of the session, and these files can degrade performance as they grow in size.1
• We formulate context curation as a capacitated assortment problem.1
• We prove an upper bound on the optimal file size, regardless of the number of available candidate instructions, and that appending every instruction with positive standalone value can be arbitrarily worse in net value than selecting an optimal subset.1
• Empirical experiments further show that irrelevant rules drawn from real context files reduce language-model compliance.1

Sources (1)

• [1]Curating Always-Loaded Context for LLM Agents: A Capacitated Assortment Model with Censored Feedback
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 7, 11:44 PM

  At the start of every session, LLM agents load a fixed context file, such as $AGENTS.md$.

  Each loaded token in the file is charged again in every later round of the session, and these files can degrade performance as they grow in size.

Extractive summary: sentences quoted from the sources.