Tag
This paper investigates holographic reduced representations for zero-shot compositional reasoning in knowledge graphs, finding that while single-hop performance is strong, composition fails due to retrieval capacity and interference effects in the superposed memory, not the bind-unbind algebra.
Diverge-to-Induce Prompting (DIP) enhances zero-shot reasoning by first generating multiple diverse rationales, elaborating each into a detailed plan, and then inducing a final plan, outperforming single-strategy prompting methods.
TRN-R1-Zero introduces a post-training framework that enables LLMs to perform zero-shot reasoning on text-rich networks using only reinforcement learning, without supervised fine-tuning or chain-of-thought data.