Instruction Capacity: Behavior and Degradation of Language Models under Controlled Load
arXiv:2610.36095v1 Announce Type: new Abstract: The quality with which language models follow instructions falls as the load grows, yet there is still no quantity that expresses the limit of a model in units of load with an uncertainty. We propose an instrument in which the load consists of rules t…
Read the full story at arXiv cs.LG ↗
Timeline · 1 report
- 2026-10-28 04:00 · arXiv cs.LG
Instruction Capacity: Behavior and Degradation of Language Models under Controlled Load