Get started
Author

Anton Korznikov

Sanity Checks for Sparse Autoencoders: Do SAEs Beat Random Baselines?
2026 · Korznikov, Anton, Galichin, Andrey, et al. · arXiv (Cornell University) · 1
PDF
Context or Knowledge: How LLMs Attribute Their Own Answers
2026 · Afendulev, Kirill, Dontsov, Alexey, et al. · arXiv (Cornell University)
PDF
From Stories to Features: Scaling Concept Vectors for AI Alignment & Interpretability
2026 · Scherbakov, Dmitry, Chernikov, Oleg, et al. · arXiv (Cornell University)
PDF
Sparse Transformer: a Controlled Testbed for Mechanistic Interpretability
2026 · Korznikov, Anton, Dontsov, Alexey, et al. · arXiv (Cornell University)
PDF
Your Transformer Can Hold Two Thoughts at Once: Evidence of Linear Superposition in LLMs
2026 · Tikhonov, Pavel, Korznikov, Anton, et al. · arXiv (Cornell University)
PDF
Feature Drift: How Fine-Tuning Repurposes Representations in LLMs
2026 · Association for Computational Linguistics 2026, Dontsov, Alexey, et al. · Underline Science Inc.
Out of Distribution, Out of Luck: Process Rewards Misguide Reasoning Models
2026 · Alexey Dontsov, Anton Korznikov, et al. · Conference of the European Chapter of the Association for Computational Linguistics
PDF
Feature Drift: How Fine-Tuning Repurposes Representations in LLMs
2026 · Andrey V. Galichin, Anton Korznikov, et al. · Findings of the Association for Computational Linguistics: EACL
PDF
What Attention Recalls and Recurrence Controls in Hybrid Language Models
2026 · Kirill Afendulev, Alexey Dontsov, et al. · Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP)
PDF
SmurfCat at SHROOM-CAP: Factual but Awkward? Fluent but Wrong? Tackling Both in LLM Scientific QA
2025 · Timur Ionov, Evgenii Nikolaev, et al.
PDF
Works count: 12
Citation count: 1
H-index: 1
I10-index:
Year
Topic
Type