A comprehensive index of artificial intelligence and machine-learning research with AI-generated summaries, citation metrics, and direct links to papers and code.
Canbin Huang, Tianyuan Shi, Yuhua Zhu, et al.
Lookahead routing predicts latent representations of potential model outputs to guide LLM selection, improving routing decisions without full inference.
Arian Hosseini, Alessandro Sordoni, Daniel Toyama, et al.
This paper reveals a significant reasoning gap in LLMs when solving compositional math problems, showing that performance on standard benchmarks masks systematic differences in reasoning abilities.
Vatsal Venkatkrishna, Nico Daheim, Iryna Gurevych
Introduces parameter-space exploration for LLM RL via Perturbed Parameter Policy Optimization (3PO), improving downstream performance over GRPO at similar FLOPs.
Tianyu Huai, Tingshuo Fan, Xinchi Chen, et al.
AgentHPOBench is a sequential benchmark with 30 executable ML tasks evaluating LLM agents' ability to iteratively optimize hyperparameters based on experimental evidence.
Yifan Ding, Xincheng Wei, Yoshua Y. Li, et al.
SAF introduces a stable advantage fusion framework for on-policy distillation, resolving magnitude and temporal mismatches to improve RLVR+OPD training.
Manuel Cossio
This paper provides a comprehensive taxonomy of LLM hallucinations, arguing their theoretical inevitability and emphasizing the need for robust detection, mitigation, and human oversight.
Maanak Gupta, Charankumar Akiri, Kshitiz Aryal, et al.
This paper explores the dual-use of generative AI in cybersecurity, demonstrating attack techniques like jailbreaks and prompt injection, while also proposing defensive applications.
Krzysztof Czarnecki, Ulrich W. Eisenecker
This book introduces generative programming, a paradigm for automating software component assembly using domain engineering, feature modeling, and code generation.
Pranjal Kumar
This survey provides a comprehensive overview of large language models, covering their technical frameworks, applications across domains, and future challenges.
Andrea Matarazzo, Riccardo Torlone
A survey exploring foundational components, scaling mechanisms, and emergent abilities of LLMs, including CoT and PoT reasoning, with applications across sectors.
Unknown
AceCoder improves code generation by using automated large-scale test-case synthesis to train reward models for reinforcement learning.
Unknown
Instructflow uses adaptive symbolic constraints to guide code generation for long-horizon robotic planning.