SV
Sohan Venkatesh
cs.CLcs.LGcs.AI
On Valency
published · living versionsW_4d9v7zp6·v1 · currentpublished
Knowingly Wrong: Training LLMs Against Their Own Beliefs can Induce Misalignment
with Ariana M. Villegas-Suarez, Guido Freire, Bartosz Cywiński
1 version
Preprints & journals
5 papers in the corpus · 2026Repeated-Token Counting Reveals a Dissociation Between Representations and Outputs2605.09239v2 · Sohan Venkatesh2026 · 0 citationsarXiv
Architecture, Not Scale: Circuit Localization in Large Language Models2605.08853v1 · Sohan Venkatesh2026 · 0 citationsarXiv
Negative Before Positive: Asymmetric Valence Processing in Large Language Models2605.05653v1 · Sohan Venkatesh2026 · 0 citationsarXiv
Algorithmic Blindness in Large Language Models: A Calibration Study of Performance Prediction2602.21947v5 · Sohan Venkatesh, Ashish Mahendran Kurapath, Tejas Melkote2026 · 0 citationsarXiv
On the Non-Identifiability of Steering Vectors in Large Language Models2602.06801v4 · Sohan Venkatesh, Ashish Mahendran Kurapath2026 · 0 citationsarXiv
Career total: 5 works. 5 are in this corpus.
Profile built from the corpus for this byline.
Author records are still filling in while the Hub is in alpha. If this is your page, you'll be able to claim it soon. Spot a mistake? Tell us.