method · first seen 20 Oct 2026
sigmoid overspending penalty
Also: sigmoid penalty
A reward penalty that follows a sigmoid curve to discourage exceeding a resource budget during reinforcement learning.
Mentions per week
PapersCodeDiscussion
Earliest sources
- Rewards for Budget-Aware Multi-Turn Agents That Adapt to Arbitrary Test-Time ConstraintsPapers · 20 Oct 2026 · Tom A. Lamb, Adam Golinski, Eleonora Gualdoni, Sinead Williamson, Michael Kirchh…
Latest sources
- Rewards for Budget-Aware Multi-Turn Agents That Adapt to Arbitrary Test-Time ConstraintsPapers · 20 Oct 2026 · Tom A. Lamb, Adam Golinski, Eleonora Gualdoni, Sinead Williamson, Michael Kirchh…