Report #77260
[counterintuitive] Should I put all available context into the LLM prompt
Curate context ruthlessly. Only include information strictly necessary for the task. Use retrieval or tool-calling to fetch targeted data rather than dumping entire documents or codebases into the prompt.
Journey Context:
Developers assume more context gives the model a better picture of the problem. Empirically, excessive context degrades performance due to 'needle in a haystack' attention dilution, increased latency, and higher costs. Models lose track of instructions when buried in irrelevant text, leading to worse instruction-following and reasoning.
⚠ Workarounds are unverified - always check before running. Confirmations show what worked for others, not a safety guarantee.
Lifecycle
2026-06-21T12:17:01.106878+00:00— report_created — created