Related papers: Recite, Reconstruct, Recollect: Memorization in LM…

In computer simulation of the learning process is usually assumed that all elements of the training material are assimilated equally durable. But in practice, the knowledge, which a student uses in its operations, are remembered much…

When language models are trained on textual data, they acquire both knowledge about the structure of language as well as knowledge of facts about the world. At inference time, their knowledge of facts can be leveraged to solve interesting…

Most language model pre-training frameworks concatenate multiple documents into fixed-length sequences and use causal masking to compute the likelihood of each token given its context; this strategy is widely adopted due to its simplicity…

Memory is inherently entangled with prediction and planning. Flexible behavior in biological and artificial agents depends on the interplay of learning from the past and predicting the future in ever-changing environments. This chapter…

Answering multi-hop reasoning questions requires retrieving and synthesizing information from diverse sources. Language models (LMs) struggle to perform such reasoning consistently. We propose an approach to pinpoint and rectify multi-hop…

While fine-tuning is the standard for injecting factual knowledge into large language models (LLMs), the mechanisms enabling reliable fact recall via unseen queries remain poorly understood. Common two-stage training strategies, which…

We study how large language models recall relational knowledge during text generation, with a focus on identifying latent representations suitable for relation classification via linear probes. Prior work shows how attention heads and MLPs…

Diffusion models are a class of generative models that learn to synthesize samples by inverting a diffusion process that gradually maps data into noise. While these models have enjoyed great success recently, a full theoretical…

Subword tokenization introduces a computational layer in language models where many distinct token sequences decode to the same surface form and preserve meaning, yet induce different internal computations. Despite this non-uniqueness,…

Memorization impacts the performance of deep learning algorithms. Prior works have studied memorization primarily in the context of generalization and privacy. This work studies the memorization effect on incremental learning scenarios.…

Evaluation of multimodal reasoning models is typically reduced to a single accuracy score, implicitly treating reasoning as a unitary capability. We introduce MathLens, a benchmark of textbook-style geometry problems that exposes this…

Are the predictions of humans and language models affected by similar things? Research suggests that while comprehending language, humans make predictions about upcoming words, with more predictable words being processed more easily.…

Memory-augmented large language models extend reasoning beyond a fixed context window by maintaining long-term memory across interactions. However, existing memory systems often collapse stable user facts, episodic events, and behavioral…

Transformer models are now a cornerstone in natural language processing. Yet, explaining their decisions remains a challenge. It was shown recently that the same model trained on the same data with a different randomness can lead to very…

Training data plays a pivotal role in AI models. Large language models (LLMs) are trained with massive amounts of documents, and their parameters hold document-related contents. Recently, several studies identified content-specific…

Despite their success in image generation, diffusion models can memorize training data, raising serious privacy and copyright concerns. Although prior work has sought to characterize, detect, and mitigate memorization, the fundamental…

When a language model is trained to predict natural language sequences, its prediction at each moment depends on a representation of prior context. What kind of information about the prior context can language models retrieve? We tested…

Semiparametric language models (LMs) have shown promise in continuously learning from new text data by combining a parameterized neural LM with a growable non-parametric memory for memorizing new content. However, conventional…

Despite the increasing prevalence of large language models (LLMs), we still have a limited understanding of how their representational spaces are structured. This limits our ability to interpret how and what they learn or relate them to…

Autoregressive language models are powerful and relatively easy to train. However, these models are usually trained without explicit conditioning labels and do not offer easy ways to control global aspects such as sentiment or topic during…