arXiv:2605. 25848v2 Announce Type: replace-cross Abstract: A concept probe is only as reliable as the layer it is taken from.
By James Henry
arXiv:2608. 10251v1 Announce Type: cross Abstract: A transformer's answer lives on one axis: the direction its unembedding reads.
By Mark Oskin
arXiv:2607. 10578v1 Announce Type: new Abstract: Existing hypotheses represent a concept in an LLM as a single point, a linear direction, or a Gaussian cluster, yet it remains unclear how and why such structures emerge.
By Chunwei Ma, Russell Wolfinger
arXiv:2606. 12629v1 Announce Type: cross Abstract: We show that the standard basis of transformer hidden states already provides a training-free, architecture-general feature basis.
By Varun Reddy Nalagatla
arXiv:2609.37824v1 Announce Type: cross
Abstract: Transformer language models build predictions through successive residual updates, but how their representations become specific to an eventual outco...
By Timur Mudarisov, Mikhail Burtsev, Radu State
The paper introduces Exemplar Partitioning (EP), an unsupervised technique that constructs interpretable feature dictionaries from large language model activations by clustering streamed activations into Voronoi regions defined by exemplars and their averages. EP allows comparison of dictionaries across layers, checkpoints, and architectures, and demonstrates utility in interpreting model behavior, tracking training dynamics, detecting hidden concepts, and enabling targeted interventions. Experiments on Gemma‑2‑2B and Llama‑3.1‑8B show EP can reveal how instruction tuning reorganizes harmful prompt activations, facilitate interventions that alter model responses, and achieve high concept‑detection performance while requiring far fewer construction tokens than comparable methods.
By Jessica Rumbelow