RESEARCH · RESEARCH · #1768
arXiv:2610.11007v1 models LLM always-loaded context as a capacitated-assortment problem with censored feedback
This paper formulates the problem of curating fixed per-session context files (e.g., AGENTS.md) for LLM agents as a capacitated-assortment model where instructions consume finite attention (token) capacity and retained instructions incur per-session setup cost. The authors prove an upper bound on optimal file size, show appending every positively valued instruction can be arbitrarily suboptimal, analyze learning under censored feedback (missing instructions only produce signal when they cause harm), provide bounds on required evidence and regret with limited human reviewers, and report experiments showing irrelevant rules in real context files reduce model compliance.
KEY POINTS
- This paper formulates the problem of curating fixed per-session context files (e.g., AGENTS.md) for LLM agents as a capacitated-assortment model where instructions consume finite attention (token) capacity and retained instructions incur per-session setup cost.
- The authors prove an upper bound on optimal file size, show appending every positively valued instruction can be arbitrarily suboptimal, analyze learning under censored feedback (missing instructions only produce signal when they cause harm), provide bounds on required evidence and regret with limited human reviewers, and report experiments showing irrelevant rules in real context files reduce model compliance.
- Understanding and bounding the trade-offs of always-loaded context (size limits, censored feedback, and edit constraints) matters for reliable LLM agent behavior and scalable context curation.
WHY IT MATTERS
Understanding and bounding the trade-offs of always-loaded context (size limits, censored feedback, and edit constraints) matters for reliable LLM agent behavior and scalable context curation.