Proactively Detecting Threats: A Novel Approach Using LLMs

Aniesh Chawla 1 , Udbhav Prasad 1 · 2026

Read Proactively Detecting Threats: A Novel Approach Using LLMs by Aniesh Chawla 1 , Udbhav Prasad 1 (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-09029.txt.

Description

Aniesh Chawla and Udbhav Prasad test chat models as early warn scouts for cyber threats. Most bad code hunts are reactive: they wait for a known bad file. A scout reads loose web text on new threats and pulls out signs of risk, such as bad hosts or file prints. Such signs let a firm block a threat before it hits.

The pair give the first broad test of chat models on this scout task. They check if models can pull true signs from noise and hype. False signs waste staff time, so care counts as much as reach. A reader in cyber work gets a first scale for AI scouts. It shows if they can aid a guard team, or just add noise.

Catalog record

CREATORAniesh Chawla 1 , Udbhav Prasad 1
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE16 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2601-09029

Browse the full library · About the Tower of Babel