Related papers: Position: Stop Anthropomorphizing Intermediate Tok…

We study the interpretability issue of task-oriented dialogue systems in this paper. Previously, most neural-based task-oriented dialogue systems employ an implicit reasoning strategy that makes the model predictions uninterpretable to…

Token-based transformer world models have shown strong performance in visual reinforcement learning, but often suffer from temporal inconsistency in long-horizon rollouts, including object duplication, disappearance, and transmutation. A…

Conceptual reasoning, the ability to reason in abstract and high-level perspectives, is key to generalization in human cognition. However, limited study has been done on large language models' capability to perform conceptual reasoning. In…

In recent years, the detection of AI-generated text has become a critical area of research due to concerns about academic integrity, misinformation, and ethical AI deployment. This paper presents COT Fine-tuned, a novel framework for…

Crosslingual conditional generation (e.g., machine translation) has long enjoyed the benefits of scaling. Nonetheless, there are still issues that scale alone may not overcome. A source query in one language, for instance, may yield several…

Language models (LMs) are increasingly extended with new learnable vocabulary tokens for domain-specific tasks, such as Semantic-ID tokens in generative recommendation. The standard practice initializes these new tokens as the mean of…

We present a framework where neural models develop an AI Mother Tongue, a native symbolic language that simultaneously supports intuitive reasoning, compositional symbol chains, and inherent interpretability. Unlike post-hoc explanation…

We study privacy leakage in the reasoning traces of large reasoning models used as personal agents. Unlike final outputs, reasoning traces are often assumed to be internal and safe. We challenge this assumption by showing that reasoning…

Getting language models to reason correctly about code requires training on data where each reasoning step can be checked. Current synthetic Chain-of-Thought (CoT) training data often consists of plausible-sounding explanations generated by…

Language models trained via outcome-based reinforcement learning (RL) to reason using chain-of-thought (CoT) have shown remarkable performance. Monitoring such a model's CoT may allow us to understand its intentions and detect potential…

LLMorphism is the biased belief that human cognition works like a large language model. I argue that the rise of conversational LLMs may make this bias increasingly psychologically available. When artificial systems produce human-like…

Large language models increasingly rely on long chains of thought to improve accuracy, yet such gains come with substantial inference-time costs. We revisit token-efficient post-training and argue that existing sequence-level reward-shaping…

Many educational technologies use artificial intelligence (AI) that presents generated or produced language to the learner. We contend that all language, including all AI communication, encodes information about the identity of the human or…

Latent tokens are gaining attention for enhancing reasoning in large language models (LLMs), yet their internal mechanisms remain unclear. This paper examines the problem from a reliability perspective, uncovering fundamental weaknesses:…

When humans solve complex problems, they typically create a sequence of ideas (involving an intuitive decision, reflection, error correction, etc.) in order to reach a conclusive decision. Contrary to this, today's models are mostly trained…

Language models lack the notion of interchangeable tokens: symbols that are semantically equivalent yet distinct, such as bound variables in formal logic. This limitation prevents generalization to larger vocabularies and hinders the…

We introduce Transformer Grammars (TGs), a novel class of Transformer language models that combine (i) the expressive power, scalability, and strong performance of Transformers and (ii) recursive syntactic compositions, which here are…

The chain-of-thought technique has been received well in multi-modal tasks. It is a step-by-step linear reasoning process that adjusts the length of the chain to improve the performance of generated prompts. However, human thought processes…

Retrieval-Augmented Generation (RAG) utilizes external knowledge to augment Large Language Models' (LLMs) reliability. For flexibility, agentic RAG employs autonomous, multi-round retrieval and reasoning to resolve queries. Although recent…

Latent reasoning via continuous chain-of-thoughts (Latent CoT) has emerged as a promising alternative to discrete CoT reasoning. Operating in continuous space increases expressivity and has been hypothesized to enable superposition: the…