Interpretability provides a toolset for understanding how and why language models behave in certain ways. However, there is little unity in the field: Most studies use ad-hoc evaluations and do not share theoretical foundations, making it difficult to measure progress and compare the pros and cons of different techniques. Furthermore, while mechanistic understanding is frequently discussed, the basic causal units underlying these mechanisms are often not explicitly defined. In this article, we propose a perspective on interpretability research grounded in causal mediation analysis. Specifically, we describe the history and current state of interpretability taxonomized according to the types of causal units (mediators) utilized, as well as methods used to search over mediators. We discuss the pros and cons of each mediator, providing insights as to when particular kinds of mediators and search methods are most appropriate. We argue that this framing yields a more cohesive narrative of the field and helps researchers select appropriate methods based on their research objective. Our analysis yields actionable recommendations for future work, including the discovery of new mediators and the development of standardized evaluations tailored to these goals.

We know from prior work that LLMs encode social biases, and that this manifests in clinical tasks. In this work we adopt tools from mechanistic interpretability to unveil sociodemographic representations and biases within LLMs in the context of healthcare. Specifically, we ask: Can we identify activations within LLMs that encode sociodemographic information (e.g., gender, race)? We find that, in three open weight LLMs, gender information is highly localized in MLP layers and can be reliably manipulated at inference time via patching. Such interventions can surgically alter generated clinical vignettes for specific conditions, and also influence downstream clinical predictions which correlate with gender, e.g., patient risk of depression. We find that representation of patient race is somewhat more distributed, but can also be intervened upon, to a degree. To our knowledge, this is the first application of mechanistic interpretability methods to LLMs for healthcare.

This paper presents a computational approach for creating a dataset on communal violence in the context of Bangladesh and West Bengal of India and benchmark evaluation. In recent years, social media has been used as a weapon by factions of different religions and backgrounds to incite hatred, resulting in physical communal violence and causing death and destruction. To prevent such abusive use of online platforms, we propose a framework for classifying online posts using an adaptive question-based approach. We collected more than 168,000 YouTube comments from a set of manually selected videos known for inciting violence in Bangladesh and West Bengal. Using both unsupervised and later semi-supervised topic modeling methods on those unstructured data, we discovered the major word clusters to interpret the related topics of peace and violence. Topic words were later used to select 20,142 posts related to peace and violence of which we annotated a total of 6,046 posts. Finally, we applied different modeling techniques based on linguistic features, and sentence transformers to benchmark the labeled dataset with the best-performing model reaching ~71% macro F1 score.

Social media platforms and online streaming services have spawned a new breed of Hate Speech (HS). Due to the massive amount of user-generated content on these sites, modern machine learning techniques are found to be feasible and cost-effective to tackle this problem. However, linguistically diverse datasets covering different social contexts in which offensive language is typically used are required to train generalizable models. In this paper, we identify the shortcomings of existing Bangla HS datasets and introduce a large manually labeled dataset BD-SHS that includes HS in different social contexts. The labeling criteria were prepared following a hierarchical annotation process, which is the first of its kind in Bangla HS to the best of our knowledge. The dataset includes more than 50,200 offensive comments crawled from online social networking sites and is at least 60% larger than any existing Bangla HS datasets. We present the benchmark result of our dataset by training different NLP models resulting in the best one achieving an F1-score of 91.0%. In our experiments, we found that a word embedding trained exclusively using 1.47 million comments from social media and streaming sites consistently resulted in better modeling of HS detection in comparison to other pre-trained embeddings. Our dataset and all accompanying codes is publicly available at github.com/naurosromim/hate-speech-dataset-for-Bengali-social-media