Lovkush Agarwal

activity

20142024

most citedExtracting Paragraphs from LLM Token Activations

2 citations · 5 across the 3 of their papers we have counts for

collaborators

3 papers

cs.CL2024★ 2 cited

Extracting Paragraphs from LLM Token Activations

Nicholas Pochinkov, Angelo Benoit, Lovkush Agarwal +2

Generative large language models (LLMs) excel in natural language processing tasks, yet their inner workings remain underexplored beyond token-level predictions. This study investi…

math.LO2015★ 2 cited

$2^{ℵ_{0}}$ pairwise non-isomorphic maximal-closed subgroups of Sym $(N)$ via the classification of the reducts of the Henson digraphs

Lovkush Agarwal, Michael Kompatscher

Given two structures $M$ and $N$ on the same domain, we say that $N$ is a reduct of $M$ if all $\emptyset$ -definable relations of $\mathcal{…

math.LO2014★ 1 cited

Reducts of the Generic Digraph

Lovkush Agarwal

The generic digraph $(D, E)$ is the unique countable homogeneous digraph that embeds all finite digraphs. In this paper, we determine the lattice of reducts of $(D, E)$ , where a stru…

Extracting Paragraphs from LLM Token Activations

2ℵ0​ pairwise non-isomorphic maximal-closed subgroups of Sym(N) via the classification of the reducts of the Henson digraphs

Reducts of the Generic Digraph

$2^{ℵ_{0}}$ pairwise non-isomorphic maximal-closed subgroups of Sym $(N)$ via the classification of the reducts of the Henson digraphs