Understanding and Controlling a Maze-Solving Policy Network

Autor:	Mini, Ulisse, Grietzer, Peli, Sharma, Mrinank, Meek, Austin, MacDiarmid, Monte, Turner, Alexander Matt
Rok vydání:	2023
Předmět:	Computer Science - Artificial Intelligence
Druh dokumentu:	Working Paper
Popis:	To understand the goals and goal representations of AI systems, we carefully study a pretrained reinforcement learning policy that solves mazes by navigating to a range of target squares. We find this network pursues multiple context-dependent goals, and we further identify circuits within the network that correspond to one of these goals. In particular, we identified eleven channels that track the location of the goal. By modifying these channels, either with hand-designed interventions or by combining forward passes, we can partially control the policy. We show that this network contains redundant, distributed, and retargetable goal representations, shedding light on the nature of goal-direction in trained policy networks. Comment: 46 pages
Databáze:	arXiv
Externí odkaz:	http://arxiv.org/abs/2310.08043 Zobrazit plný text záznamu View this record from Arxiv