Large Language Models (LLMs) have exhibited remarkable reasoning capabilities, achieving impressive results across a wide range of tasks. Despite these advances, significant reasoning failures persist, occurring even in seemingly simple scenarios. To systematically understand and address these shortcomings, the authors of the paper present the first comprehensive survey dedicated to reasoning failures in LLMs.
The authors introduce a novel categorization framework that distinguishes reasoning into embodied and non-embodied types, with the latter further subdivided into informal (intuitive) and formal (logical) reasoning. In parallel, the authors classify reasoning failures along a complementary axis into three types: fundamental failures intrinsic to LLM architectures that broadly affect downstream tasks; application-specific limitations that manifest in particular domains; and robustness issues characterized by inconsistent performance across minor variations. For each reasoning failure, the authors provide a clear definition, analyze existing studies, explore root causes, and present mitigation strategies.
Read more…
Source: ARXIV, Cornell University
Sign up for the Cyber Security Review Newsletter
The latest cyber security news and insights delivered right to your inbox
Related:
- Industry backs Norway’s AI powerhouse project
October 2, 2018
Norway is stepping up its efforts in the rapidly evolving and increasingly commercial artificial intelligence (AI) sector. Participants in the latest programme want to reduce the country’s reliance on oil revenues by tapping into AI opportunities in its growing technology sector. Norway has traditionally played a junior role in the context of Nordic technology development, investment and ...
- Researchers Developed Artificial Intelligence-Powered Stealthy Malware
August 9, 2018
Artificial Intelligence (AI) has been seen as a potential solution for automatically detecting and combating malware, and stop cyber attacks before they affect any organization. However, the same technology can also be weaponized by threat actors to power a new generation of malware that can evade even the best cyber-security defenses and infects a computer network ...
- Bot and drone misuse could lead to cybercrime explosion
February 21, 2018
The rapid development of drones and artificial intelligence is a “game-changer” that will present a serious threat to national security if it isn’t addressed. The assessment, made by 26 experts from institutions including Cambridge and Oxford Universities, warns of the potential for malicious use of artificial intelligence (AI) by rogue states, criminals, and terrorists. The panel forecast ...