A comprehensive index of artificial intelligence and machine-learning research with AI-generated summaries, citation metrics, and direct links to papers and code.
Haotong Yang, Yi Hu, Shijia Kang, et al.
This paper introduces a comprehensive benchmark for evaluating numerical understanding and processing abilities of LLMs, revealing frequent failures and the ineffectiveness of existing enhancement techniques.
Siwei Wu, Z. Peng, Xinrun Du, et al.
This paper analyzes OpenAI's o1 model's reasoning patterns, comparing it with existing test-time compute methods on general reasoning benchmarks and summarizing six distinct reasoning patterns.
Xiaofeng Wu, Alan Ritter, Wei Xu
This paper introduces a taxonomy of tabular input representations and table understanding tasks, highlighting critical gaps including retrieval-focused tasks, challenges with complex structures, and limited generalization.
Yizhou Fan, Luzhen Tang, Huixiao Le, et al.
A randomized experiment shows ChatGPT improves short-term writing performance but may induce metacognitive laziness, with no gains in intrinsic motivation or knowledge transfer.
Armin Steinhauser
TinyCast is an attention-free zero-shot forecaster with 146K parameters that computes periodicity via a spectral detector, achieving state-of-the-art probabilistic accuracy at its size.
Xiaomin Li, Zhou Yu, Zhiwei Zhang, et al.
This paper reveals that chain-of-thought reasoning can significantly degrade instruction-following accuracy in LLMs, and proposes selective reasoning strategies to mitigate the issue.
Rujikorn Charakorn, Edoardo Cetin, Shinnosuke Uesaka, et al.
Doc-to-LoRA (D2L) is a lightweight hypernetwork that meta-learns to perform approximate context distillation in a single forward pass, generating LoRA adapters for a target LLM to reduce inference latency and KV-cache memory.
Tianyu Wu, Shizhu He, Jingping Liu, et al.
This paper provides a concise overview of ChatGPT's history, core techniques, pros and cons, and potential future development, serving as an entry point for understanding the model.
Michael Hoy, Alexey S. Matveev, Andrey V. Savkin
This survey reviews collision-free navigation techniques for unmanned vehicles in unknown cluttered environments, focusing on reactive local-sensory-data methods and recent model predictive and sliding mode control laws.
Mathura Shanmugasundaram, Arunkumar Tamilarasu
This review examines the dual positive and negative impacts of digital technology, social media, and AI on cognitive functions across age groups, proposing strategies to optimize digital engagement.
Liam Barkley, Brink van der Merwe
This paper empirically evaluates prompting strategies and tool-calling agents to reduce hallucinations in LLMs, finding that simpler techniques often outperform complex ones and that tool use can increase hallucination rates.
Yuanhe Zhang, Weiliu Wang, Jie Ren, et al.
Proposes ILL, an inaudible red teaming method using low-frequency waveforms to attack LALMs, and DRG, a defense that detects distribution shifts and requests reacquisition, reducing accuracy by up to 67 points.