Multi-step reasoning is widely adopted in the community to explore the better performance of language models (LMs). We report on the systematic strategy that LMs use in this process. Our controlled experiments reveal that LMs rely more heavily on heuristics, such as lexical overlap, in the earlier stages of reasoning Figure 1: Illustration of the systematic strategy we discovered when more steps are required to reach an in language models (LMs). When the goal is answer. Conversely, as LMs progress closer distant from the current state in a multi-step reasoning to the final answer, their reliance on heuristics process, the models tend to rely on heuristics, such as decreases. This suggests that LMs track only superficial overlap, which can lead them in the wrong a limited number of future steps and dynamically direction. In contrast, when the goal is within a limited combine heuristic strategies with logical distance, the models are more likely to take rational actions ones in tasks involving multi-step reasoning.

The Peggy Smedley Show is the most influential IoT (Internet of Things) podcasts engaging, educating, and entertaining listeners about relevant digital tech trends. Tune in to the live show Tuesday at 12 p.m. CT or listen to podcasts downloadable through iHeartRadio and iTunes. Peggy Smedley is joined by Josh Peschel, assistant professor of agriculture and Biosystems engineering, Iowa State University, who says there isn't a one-size-fits-all solution to all of the problems geographically--there are too many variables. Peggy kicks off a new month, looking at the topic of infrastructure. Peggy Smedley is joined by Juan Pablo Segura, cofounder, Babyscripts, who talks about the demand that is being fueled on the patient side in healthcare.