Related papers: A Measure-Theoretic Characterization of Tight Lang…
Large language models (LLMs) are susceptible to memorizing training data, raising concerns about the potential extraction of sensitive information at generation time. Discoverable extraction is the most common method for measuring this…
A controversial test for Large Language Models concerns the ability to discern possible from impossible language. While some evidence attests to the models' sensitivity to what crosses the limits of grammatically impossible language, this…
The entropy rate of printed English is famously estimated to be about one bit per character, a benchmark that modern large language models (LLMs) have only recently approached. This entropy rate implies that English contains nearly 80…
Some consider large-scale language models that can generate long and coherent pieces of text as dangerous, since they may be used in misinformation campaigns. Here we formulate large-scale language model output detection as a hypothesis…
Modern language models define distributions over strings, but downstream tasks often require different output formats. For instance, a model that generates byte-pair strings does not directly produce word-level predictions, and a DNA model…
Semantic leakage is a phenomenon recently introduced by Gonen et al. (2024). It refers to a situation in which associations learnt from the training data emerge in language model generations in an unexpected and sometimes undesired way.…
Language modeling studies the probability distributions over strings of texts. It is one of the most fundamental tasks in natural language processing (NLP). It has been widely used in text generation, speech recognition, machine…
Spoken communication occurs in a "noisy channel" characterized by high levels of environmental noise, variability within and between speakers, and lexical and syntactic ambiguity. Given these properties of the received linguistic input,…
Pointwise maximal leakage (PML) is an operationally meaningful privacy measure that quantifies the amount of information leaking about a secret $X$ to a single outcome of a related random variable $Y$. In this paper, we extend the notion of…
In many current speech recognizers, a statistical language model is used to indicate how likely it is that a certain word will be spoken next, given the words recognized so far. How can statistical language models be improved so that more…
Building on research arguing for the possibility of conceptual and categorical knowledge acquisition through statistics contained in language, we evaluate predictive language models (LMs) -- informed solely by textual input -- on a…
We consider the problem of inferring the probability distribution associated with a language, given data consisting of an infinite sequence of elements of the languge. We do this under two assumptions on the algorithms concerned: (i) like a…
By positing a relationship between naturalistic reading times and information-theoretic surprisal, surprisal theory (Hale, 2001; Levy, 2008) provides a natural interface between language models and psycholinguistic models. This paper…
We propose an alternate approach to quantifying how well language models learn natural language: we ask how well they match the statistical tendencies of natural language. To answer this question, we analyze whether text generated from…
Large Language Models (LLMs) are trained on massive web-crawled corpora. This poses risks of leakage, including personal information, copyrighted texts, and benchmark datasets. Such leakage leads to undermining human trust in AI due to…
How predictable a word is can be quantified in two ways: using human responses to the cloze task or using probabilities from language models (LMs).When used as predictors of processing effort, LM probabilities outperform probabilities…
Data leakage has been identified in 648 published machine learning papers across 30 scientific fields. The knowledge to prevent it exists; the tools do not enforce it. This paper presents a grammar - eight typed primitives, a directed…
Language modeling has shifted in recent years from a distribution over strings to prediction models with textual inputs and outputs for general-purpose tasks. This position paper highlights the often overlooked implications of this shift…
This work quantifies the risk of training data leakage from LLMs (Large Language Models) using sequence-level probabilities. Computing extraction probabilities for individual sequences provides finer-grained information than has been…
Language models typically tokenize raw text into sequences of subword identifiers from a predefined vocabulary, a process inherently sensitive to typographical errors, length variations, and largely oblivious to the internal structure of…