Related papers: Triangulation as an Acceptance Rule for Multilingu…

Most of recent work in cross-lingual word embeddings is severely Anglocentric. The vast majority of lexicon induction evaluation dictionaries are between English and another language, and the English embedding space is selected by default…

Causal abstraction provides a theoretical foundation for mechanistic interpretability, the field concerned with providing intelligible algorithms that are faithful simplifications of the known, but opaque low-level details of black box AI…

Mechanistic interpretability aims to break models into meaningful parts; verifying that two such parts implement the same computation is a prerequisite. Existing similarity measures evaluate either empirical behaviour, leaving them blind to…

Morphological inflection is a popular task in sub-word NLP with both practical and cognitive applications. For years now, state-of-the-art systems have reported high, but also highly variable, performance across data sets and languages. We…

Reinforcement learning has proven effective for enhancing multi-step reasoning in large language models (LLMs), yet its benefits have not fully translated to multilingual contexts. Existing methods struggle with a fundamental trade-off:…

Language Models (LMs) have emerged as powerful sources of evidence for linguists seeking to develop theories of syntax. In this paper, we argue that causal interpretability methods, applied to LMs, can greatly enhance the value of such…

The framework of Pearl's Causal Hierarchy (PCH) formalizes three types of reasoning: probabilistic (i.e. purely observational), interventional, and counterfactual, that reflect the progressive sophistication of human thought regarding…

To demarcate the limits of experimental knowledge we probe the limits of what might be called an experiment. By appeal to examples of scientific practice from astrophysics and analogue gravity, we demonstrate that the reliability of…

With the blooming of various Pre-trained Language Models (PLMs), Machine Reading Comprehension (MRC) has embraced significant improvements on various benchmarks and even surpass human performances. However, the existing works only target on…

Machine learning methods can be a valuable aid in the scientific process, but they need to face challenging settings where data come from inhomogeneous experimental conditions. Recent meta-learning methods have made significant progress in…

Code-switching is a widespread practice among the world's multilingual majority, yet few benchmarks accurately reflect its complexity in everyday communication. We present PingPong, a benchmark for natural multi-party code-switching…

This paper investigates the reliability of explanations generated by large language models (LLMs) when prompted to explain their previous output. We evaluate two kinds of such self-explanations - extractive and counterfactual - using three…

The rise of the term "mechanistic interpretability" has accompanied increasing interest in understanding neural models -- particularly language models. However, this jargon has also led to a fair amount of confusion. So, what does it mean…

Mechanistic interpretability aims to understand how neural networks generalize beyond their training data by reverse-engineering their internal structures. We introduce patterning as the dual problem: given a desired form of generalization,…

As machine learning becomes more widespread and is used in more critical applications, it's important to provide explanations for these models, to prevent unintended behavior. Unfortunately, many current interpretability methods struggle…

Code-switching is a common phenomenon among multilingual speakers, where alternation between two or more languages occurs within the context of a single conversation. While multilingual humans can seamlessly switch back and forth between…

In this review, we examine the problem of designing interpretable and explainable machine learning models. Interpretability and explainability lie at the core of many machine learning and statistical applications in medicine, economics,…

In order to alleviate the inefficiencies caused by the interaction of the logic and functional sides, integrated languages may take advantage of \emph{demand} information -- i.e. knowing in advance which computations are needed and, to…

Pluralistic alignment is typically operationalised as preference aggregation: producing responses that span (Overton), steer toward (Steerable), or proportionally represent (Distributional) diverse human values. We argue that aggregation…

Mechanistic interpretability (MI) seeks to uncover how language models (LMs) implement specific behaviors, yet measuring progress in MI remains challenging. The recently released Mechanistic Interpretability Benchmark (MIB; Mueller et al.,…