Related papers: Transcoder Adapters for Reasoning-Model Diffing
Reasoning models leverage inference-time compute to significantly enhance the performance of language models on difficult logical tasks, and have become a dominating paradigm in frontier LLMs. Despite their wide adoption, the mechanisms…
Recent reasoning Large Language Models (LLMs) demonstrate remarkable problem-solving abilities but often generate long thinking traces whose utility is unclear. Our work aims to improve their efficiency, enabling them to reach high…
Decoding from large language models (LLMs) typically relies on fixed sampling hyperparameters (e.g., temperature, top-p), despite substantial variation in task difficulty and uncertainty across prompts and individual decoding steps. We…
Self-interpretation methods prompt language models to describe their own internal states, but remain unreliable due to hyperparameter sensitivity. We show that training lightweight adapters on interpretability artifacts, while keeping the…
Transformers evaluated in a single, fixed-depth pass are provably limited in expressive power to the constant-depth circuit class TC0. Running a Transformer autoregressively removes that ceiling -- first in next-token prediction and, more…
Alignment-tuned language models frequently suppress factual log-probabilities on politically sensitive topics despite retaining the knowledge in their hidden representations. We show that a 786K-parameter (approximately 0.02% of the base…
Large language models (LMs) are typically adapted to improve performance on new contexts (\eg text prompts that define new tasks or domains) through fine-tuning or prompting. However, there is an accuracy compute tradeoff -- fine-tuning…
Prior work has shown that a significant driver of performance in reasoning models is their ability to reason and self-correct. A distinctive marker in these reasoning traces is the token wait, which often signals reasoning behavior such as…
This study investigates the in-context learning capabilities of various decoder-only transformer-based language models with different model sizes and training data, including GPT2, SmolLM2, OpenELM, TinyLlama, Stable LM, and Gemma 2. We…
Recent studies on transformer-based language models show that they can answer questions by reasoning over knowledge provided as part of the context (i.e., in-context reasoning). However, since the available knowledge is often not filtered…
We investigate whether transformers use their depth adaptively across tasks of increasing difficulty. Using a controlled multi-hop relational reasoning task based on family stories, where difficulty is determined by the number of…
Recent progress in reasoning-oriented Large Language Models (LLMs) has been driven by introducing Chain-of-Thought (CoT) traces, where models generate intermediate reasoning traces before producing an answer. These traces, as in DeepSeek…
Transformer-based language models excel at both recall (retrieving memorized facts) and reasoning (performing multi-step inference), but whether these abilities rely on distinct internal mechanisms remains unclear. Distinguishing recall…
Self-supervised learning has emerged as a key approach for learning generic representations from speech data. Despite promising results in downstream tasks such as speech recognition, speaker verification, and emotion recognition, a…
Latent reasoning represents a new development in Transformer language models that has shown potential in compressing reasoning lengths compared to chain-of-thought reasoning. By directly passing the information-rich previous final latent…
Recent studies show that the reasoning capabilities of Large Language Models (LLMs) can be improved by applying Reinforcement Learning (RL) to question-answering (QA) tasks in areas such as math and coding. With a long context length, LLMs…
MLLMs have been successfully applied to multimodal embedding tasks, yet their generative reasoning capabilities remain underutilized. Directly incorporating chain-of-thought reasoning into embedding learning introduces two fundamental…
Looped Transformers offer a promising alternative to purely feed-forward computation by iteratively refining latent representations, improving language modeling and reasoning. Yet recurrent architectures remain unstable to train, costly to…
Transferring reasoning capabilities from larger language models to smaller ones through supervised fine-tuning often fails counterintuitively, with performance degrading despite access to high-quality teacher demonstrations. We identify…
Reinforcement learning substantially improves reasoning in large language models, but it also tends to lengthen chain of thought outputs and increase computational cost during both training and inference. Though length control methods have…