LLMs
Inference-Time Decomposition of Activations (ITDA): A Scalable Approach to Interpreting Large Language Models
ITDA LLMs SAEs Mechanistic Interpretability Representations Model DiffingPatrick Leask, Noura Al Moubayed, and Neel Nanda debut ITDA—a new mechanistic interpretability approach 100x faster to train than SAEs—at ICML'25
The Predictive Turn of Language. Reading the Future in the Vector Space
LLMs Multidimensional space Predictive turn Linguistic turn Prediction Translation Machine translation Transformer architecture CartographyPaolo Caffoni presents on the long historical arc of the predictive turn in a panel chaired by Matteo Pasquinelli
Questioning language models before and after AI
LLMs Multidimensional space Linguistic turn ChatMatteo Pasquinelli and Amira Moeding open the Chat, Token, Vector symposium
Empire Model Collapse: Information Entropy and the Toxic Implosion of Technofeudalism
Model collapse Labour LLMs Extractivism ImperialismMatteo Pasquinelli gives a talk at the Model Collapse festival in Vienna
Calendar feature geometry in GPT-2 layer 8 residual stream SAEs
LLMs SAEs GPT2 GeometryPatrick Leask, Bart Bussmann, and Neel Nanda take a close look at GPT-2’s SAE feature geometry on the AI Alignment Forum
The Resource Debate in Machine Translation and Large Language Models
LLMs Machine translation LinguisticsPaolo Caffoni presents at and chairs a workshop organised as a collaboration between the KIM research group and the AIModels project at Ca’ Foscari University of Venice.
Synthesizing Proteins on the Graphics Card. Protein Folding and the Limits of Critical AI Studies
Transformer architecture Protein folding Language models LLMsFabian Offert, Paul Kim, and Qiaoyu Cai look at Meta’s ESM-2 protein folding ’language model’, asking what kind of knowledge the transformer architecture produces?
Organism, Machine, Language, Model: The Making of the Artificial Mind between Animism and Formalism
Organism Machine Model Genealogy Language Linguistics LLMs CyberneticsMatteo Pasquinelli lays out a combined intellectual genealogy of AI
The Meaning Trap [Protokoll 13]
Meaning LLMs ChatGPTFabian Offert and Wolf Kittler engage with ChatGPT.
What is a language model? The contribution of semiotics to the studies of AI and automation
Semiotics LLMs AutomationMatteo Pasquinelli mobilises interpretative semiotics to illuminate the invention of algorithmic models.
Semiotics and Large Language Models
Semiotics LLMs interpretability structuralismThe KIM convened in a special colloquium to open a new chapter on interpretability research mobilising semiotics in light of large language models
The work of translation after ChatGPT and Large Language Models
translation machine translation LLMsPaolo Caffoni and Matteo Pasquinelli co-teach a seminar on machine translation in the KIM at HfG Karlsruhe
