analiza aiqbench
THOR introduces a brain-inspired framework using Theta-Gamma oscillations to address critical bottlenecks in multi-hop reasoning, specifically attention decay and error propagation across complex chains. By decoupling global planning from local retrieval with built-in verification mechanisms, the model demonstrates improved accuracy and robustness on standard benchmarks without relying solely on scaling parameters. This approach offers a novel architectural pathway for enhancing logical consistency in large language models as they tackle increasingly intricate inference tasks.
THOR: A Theta-Gamma Hierarchical Oscillatory Reasoning Framework for Multi-hop QA
arXiv:2607.20459v1 Announce Type: new Abstract: Multi-hop question answering requires retrieving and integrating evidence from multiple contexts. Despite the rapid progress of current research, multi-hop reasoning remains constrained by two persistent limitations: attention decay, where the model's focus on main question degrades as the reasoning chain grows, and error accumulation, where mistakes propagate across hops and compounds into final failure. Inspired by Theta-Gamma hierarchical oscillation which decouples global planning from local retrieval, enabling efficient attention transfer between hops and a verification and repair mechanism that interrupts the accumulation of errors in the wrong paths, we present THOR, a brain-inspired Theta-Gamma hierarchical oscillatory reasoning framework. Extensive comparative experiments and specific validation experiments on multi-hop QA benchmarks demonstrate that THOR improves answer accuracy and robustness while mitigating limitations, showcasing its generalization across different backbones.
Abstract:Multi-hop question answering requires retrieving and integrating evidence from multiple contexts. Despite the rapid progress of current research, multi-hop reasoning remains constrained by two persistent limitations: attention decay, where the model's focus on main question degrades as the reasoning chain grows, and error accumulation, where mistakes propagate across hops and compounds into final failure. Inspired by Theta-Gamma hierarchical oscillation which decouples global planning from local retrieval, enabling efficient attention transfer between hops and a verification and repair mechanism that interrupts the accumulation of errors in the wrong paths, we present THOR, a brain-inspired Theta-Gamma hierarchical oscillatory reasoning framework. Extensive comparative experiments and specific validation experiments on multi-hop QA benchmarks demonstrate that THOR improves answer accuracy and robustness while mitigating limitations, showcasing its generalization across different backbones.
Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
arXiv:2607.20459 [cs.CL]
(or arXiv:2607.20459v1 [cs.CL] for this version)
https://doi.org/10.48550/arXiv.2607.20459 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 224 cuvinte
Subiecte conexe
Exploding variance of means of exponentials: least-squares to the rescue – Machine Learning Research Blog
francisbach.com · 27 sept. 2026
SAGE: Symbolic Action-Gating and Editing for LLM Task Planners
arXiv · 29 sept. 2026
How to Build In-Vehicle AI Agents with NVIDIA: From Cloud to Car
NVIDIA · 5 mai 2026
MV-STRIDE: Enabling MLLMs to Master Multi-View Spatial Reasoning via Hierarchical Capability Modeling
arXiv · 9 sept. 2026