State of the Field
Current AI research is increasingly focused on enhancing the capabilities of large language models (LLMs) and their underlying architectures, particularly in areas like out-of-distribution generalization and task-oriented communication. Recent work has revealed significant limitations in LLMs' ability to generalize periodic patterns, prompting investigations into their underlying reasoning processes and the development of new evaluation frameworks. Researchers are also exploring novel normalization techniques to improve transformer stability and performance, while advancements in latent-space regularization are showing promise for optimizing preference learning from human feedback. Additionally, investigations into task-oriented communication protocols highlight both the efficiency and potential opacity of LLM interactions, raising important questions about transparency in AI systems. As these studies converge on understanding the nuances of model behavior, they aim to address commercial challenges in deploying AI solutions that require robust reasoning and effective communication in dynamic environments.
Papers
1–8 of 8Do Transformers Have the Ability for Periodicity Generalization?
Large language models (LLMs) based on the Transformer have demonstrated strong performance across diverse tasks. However, current models still exhibit substantial limitations in out-of-distribution (O...
Enhanced QKNorm normalization for neural transformers with the Lp norm
The normalization of query and key vectors is an essential part of the Transformer architecture. It ensures that learning is stable regardless of the scale of these vectors. Some normalization approac...
GPT-4o Lacks Core Features of Theory of Mind
Do Large Language Models (LLMs) possess a Theory of Mind (ToM)? Research into this question has focused on evaluating LLMs against benchmarks and found success across a range of social tasks. However,...
Investigating the Development of Task-Oriented Communication in Vision-Language Models
We investigate whether \emph{LLM-based agents} can develop task-oriented communication protocols that differ from standard natural language in collaborative reasoning tasks. Our focus is on two core p...
Latent Adversarial Regularization for Offline Preference Optimization
Learning from human feedback typically relies on preference optimization that constrains policy updates through token-level regularization. However, preference optimization for language models is part...
Controllable Information Production
Intrinsic Motivation (IM) is a paradigm for generating intelligent behavior without external utilities. The existing information-theoretic methods for IM are predominantly based on information transmi...
How Do Latent Reasoning Methods Perform Under Weak and Strong Supervision?
Latent reasoning has been recently proposed as a reasoning paradigm and performs multi-step reasoning through generating steps in the latent space instead of the textual space. This paradigm enables r...
Retrievit: In-context Retrieval Capabilities of Transformers, State Space Models, and Hybrid Architectures
Transformers excel at in-context retrieval but suffer from quadratic complexity with sequence length, while State Space Models (SSMs) offer efficient linear-time processing but have limited retrieval ...