Whether price-response curves exhibit threshold effects

Determine whether the apparently irregular responses of Haiku 4.5, DeepSeek V3.1, and Gemini 2.5 Flash to fuel prices reflect low measurement sensitivity or genuine threshold effects in their willingness to avoid driving over animals.

Background

The paper varies the fuel cost of swerving around an animal and examines how often models instead continue along the route and kill the animal. Some models show higher kill rates at higher prices but also occasional increases at lower prices, producing non-monotonic or gate-like patterns.

The authors explicitly state that the available experiments cannot distinguish between noise or insufficient sensitivity in the measurement instrument and a genuine threshold relationship between avoidance costs and harmful decisions. Additional experimental runs would be needed to resolve the issue.

References

Haiku, Deepseek and Gemini all end up at higher kill rates than when they began but also do sometimes kill more animals at lower fuel prices, this could either be low sensitivity in the instrument or a real threshold effect, but we would need more runs to establish this.

— HarvestBench: Measuring Whether LLM Agents Will Pay to Avoid Killing Animals  (2609.04444 - Brazilek et al., 3 Sep 2026) in Section 5.8, “The price curve of mercy”