8 papers
The Ghost Couple: Correlated LLM Name Priors and Their Haunting of the Web and Academic Publishing
Michał Brzozowski, Neo Christopher Chung
These names do not exist. Elena Vasquez and Marcus Chen have appeared as volcano experts, astronauts, thriller protagonists, podcast hosts, and academic co-authors across hundreds…
Reading the Finetuning Prior: Verbatim Content Recovery via Contrastive Decoding Diffing
Michał Brzozowski, Zuzanna Dubanowska, Enrico Cassano +1
Narrowly finetuned language models memorize implanted content verbatim, but auditing what a deployed model has been taught, without access to its weights or training data, remains…
Aligned Training: A Parameter-Free Method to Improve Feature Quality and Stability of Sparse Autoencoders (SAE)
Michał Brzozowski, Neo Christopher Chung
Sparse autoencoders (SAEs) are one of the main methods to interpret the inner workings of deep neural networks (DNNs), decomposing activations into higher-dimensional features. How…
Semantic Adapter Routing with Fine-Tuning Task Embeddings
Enrico Cassano, Michał Brzozowski, Paolo Mandica +2
Parameter-efficient fine-tuning (PEFT) has led to model ecosystems in which a single backbone is paired with many task-specialized adapters. Given such a library, routing aims to s…
Ablating Archetypes: The Stability of Archetypal SAEs is an Artifact of Initialization and Metric Design
Michał Brzozowski, Neo Christopher Chung
Dictionary learning with sparse autoencoders (SAEs) produces overcomplete bases from neural network activations that are often interpretable and reduces polysemanticity. However, f…
GPart: End-to-End Isometric Fine-Tuning via Global Parameter Partitioning
Paolo Mandica, Michał Brzozowski, Zuzanna Dubanowska +1
Low-rank adaptation (LoRA) has become a dominant paradigm for parameter-efficient fine-tuning (PEFT) of large-scale deep learning models. However, its bilinear parameterization ind…
Representation-based Broad Hallucination Detectors Fail to Generalize Out of Distribution
Zuzanna Dubanowska, Maciej Żelaszczyk, Michał Brzozowski +2
We critically assess the efficacy of the current SOTA in hallucination detection and find that its performance on the RAGTruth dataset is largely driven by a spurious correlation w…
Contrastive News and Social Media Linking using BERT for Articles and Tweets across Dual Platforms
Jan Piotrowski, Marek Wachnicki, Mateusz Perlik +9
X (formerly Twitter) has evolved into a contemporary agora, offering a platform for individuals to express opinions and viewpoints on current events. The majority of the topics dis…