Get started
Author

Nikita Balagansky

Linear Transformers with Learnable Kernel Functions are Better In-Context Models
2024 · Yaroslav Aksenov, Nikita Balagansky, et al. · Proceedings of the conference - Association for Computational Linguistics. Meeting · 5
PDF
PALBERT: Teaching ALBERT to Ponder
2022 · Nikita Balagansky, Daniil Gavrilov · neural information processing systems · 5
PDF
Ahead-of-Time P-Tuning
2023 · Gavrilov, Daniil, Balagansky, Nikita · arXiv (Cornell University) · 2
PDF
Steering LLM Reasoning Through Bias-Only Adaptation
2025 · Viacheslav Sinii, Alexey Gorbatovski, et al. · Conference on Empirical Methods in Natural Language Processing (EMNLP) · 1
PDF
Train One Sparse Autoencoder Across Multiple Sparsity Budgets to Preserve Interpretability and Accuracy
2025 · Nikita Balagansky, Yaroslav Aksenov, et al. · Conference on Empirical Methods in Natural Language Processing (EMNLP) · 1
PDF
Learn Your Reference Model for Real Good Alignment
2024 · Gorbatovski, Alexey, Shaposhnikov, Boris, et al. · arXiv (Cornell University) · 1
PDF
Diffusion Language Models Generation Can Be Halted Early
2023 · Vaina, Sofia Maria Lo Cicero, Balagansky, Nikita, et al. · arXiv (Cornell University) · 1
PDF
Linear Interpolation In Parameter Space is Good Enough for Fine-Tuned Language Models
2022 · Rofin, Mark, Balagansky, Nikita, et al. · arXiv (Cornell University) · 1
PDF
Classifiers are Better Experts for Controllable Text Generation
2022 · Sitdikov, Askhat, Balagansky, Nikita, et al. · arXiv (Cornell University) · 1
PDF
Weight Squeezing: Reparameterization for Extreme Compression and Fast Inference
2020 · Artem Chumachenko, Daniil Gavrilov, et al. · arXiv (Cornell University) · 1
Representation-Space MMD for Diffusion Language Models
2026 · Drobyshevskiy, Ilya, Sudakov, Ilia, et al. · arXiv (Cornell University)
PDF
Unstable Features, Reproducible Subspaces: Understanding Seed Dependence in Sparse Autoencoders
2026 · Gerasimov, Gleb, Rusalev, Timofei, et al. · arXiv (Cornell University)
PDF
Interpreting and Steering a Text-to-Speech Language Model with Sparse Autoencoders
2026 · Koriagin, Nikita, Aparin, Georgii, et al. · arXiv (Cornell University)
PDF
Trust-Region Behavior Blending for On-Policy Distillation
2026 · Plyusov, Daniil, Gorbatovski, Alexey, et al. · arXiv (Cornell University)
PDF
Next Embedding Prediction Makes World Models Stronger
2026 · Bredis, George, Balagansky, Nikita, et al. · arXiv (Cornell University)
PDF
You Do Not Fully Utilize Transformer’s Representation Capacity
2026 · Gleb Gerasimov, Yaroslav Aksenov, et al. · Conference on Empirical Methods in Natural Language Processing (EMNLP)
PDF
Kronecker Factorization Improves Efficiency and Interpretability of Sparse Autoencoders
2026 · Vadim Kurochkin, Yaroslav Aksenov, et al. · Conference on Empirical Methods in Natural Language Processing (EMNLP)
PDF
A Model with No Head and Many Thoughts
2026 · Nikita Koriagin, Yaroslav Aksenov, et al. · Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP)
PDF
Guided Star-Shaped Masked Diffusion
2025 · Meshchaninov, Viacheslav, Shibaev, Egor, et al. · arXiv (Cornell University)
PDF
VARAN: Variational Inference for Self-Supervised Speech Models Fine-Tuning on Downstream Tasks
2025 · Diatlova, Daria, Balagansky, Nikita, et al. · arXiv (Cornell University)
PDF
Teach Old SAEs New Domain Tricks with Boosting
2025 · Koriagin, Nikita, Aksenov, Yaroslav, et al. · arXiv (Cornell University)
PDF
Steering LLM Reasoning Through Bias-Only Adaptation
2025 · Sinii, Viacheslav, Gorbatovski, Alexey, et al. · arXiv (Cornell University)
PDF
Analyze Feature Flow to Enhance Interpretation and Steering in Language Models
2025 · Laptev, Daniil, Balagansky, Nikita, et al. · arXiv (Cornell University)
PDF
Mechanistic Permutability: Match Features Across Layers
2024 · Balagansky, Nikita, Maksimov, Ian, et al. · arXiv (Cornell University)
PDF
Linear Transformers with Learnable Kernel Functions are Better In-Context Models
2024 · Aksenov, Yaroslav, Balagansky, Nikita, et al. · arXiv (Cornell University)
PDF
Works count: 30
Citation count: 19
H-index: 2
I10-index:
Year
Topic
Type