Daniil Gavrilov

According to our database, Daniil Gavrilov authored at least 20 papers between 2019 and 2025.

Collaborative distances:

• Dijkstra number of five.
• Erdős number of five.

2025

Teach Old SAEs New Domain Tricks with Boosting.

Nikita Koriagin, Yaroslav Aksenov, Daniil Laptev, Gleb Gerasimov, Nikita Balagansky, Daniil Gavrilov

CoRR, July, 2025

ESSA: Evolutionary Strategies for Scalable Alignment.

Daria Korotyshova, Boris Shaposhnikov, Alexey Malakhov, Nikita Surnachev, George Bredis, Alexey Gorbatovski, Viacheslav Sinii, Daniil Gavrilov

CoRR, July, 2025

Train One Sparse Autoencoder Across Multiple Sparsity Budgets to Preserve Interpretability and Accuracy.

Nikita Balagansky, Yaroslav Aksenov, Daniil Laptev, Vadim Kurochkin, Gleb Gerasimov, Nikita Koryagin, Daniil Gavrilov

CoRR, May, 2025

Train Sparse Autoencoders Efficiently by Utilizing Features Correlation.

Vadim Kurochkin, Yaroslav Aksenov, Daniil Laptev, Daniil Gavrilov, Nikita Balagansky

CoRR, May, 2025

Steering LLM Reasoning Through Bias-Only Adaptation.

Viacheslav Sinii, Alexey Gorbatovski, Artem Cherepanov, Boris Shaposhnikov, Nikita Balagansky, Daniil Gavrilov

CoRR, May, 2025

You Do Not Fully Utilize Transformer's Representation Capacity.

Gleb Gerasimov, Yaroslav Aksenov, Nikita Balagansky, Viacheslav Sinii, Daniil Gavrilov

CoRR, February, 2025

Analyze Feature Flow to Enhance Interpretation and Steering in Language Models.

Daniil Laptev, Nikita Balagansky, Yaroslav Aksenov, Daniil Gavrilov

CoRR, February, 2025

The Differences Between Direct Alignment Algorithms are a Blur.

Alexey Gorbatovski, Boris Shaposhnikov, Viacheslav Sinii, Alexey Malakhov, Daniil Gavrilov

CoRR, February, 2025

Learn Your Reference Model for Real Good Alignment.

Alexey Gorbatovski, Boris Shaposhnikov, Alexey Malakhov, Nikita Surnachev, Yaroslav Aksenov, Ian Maksimov, Nikita Balagansky, Daniil Gavrilov

Proceedings of the Thirteenth International Conference on Learning Representations, 2025

Mechanistic Permutability: Match Features Across Layers.

Nikita Balagansky, Ian Maksimov, Daniil Gavrilov

Proceedings of the Thirteenth International Conference on Learning Representations, 2025

2024

Linear Transformers with Learnable Kernel Functions are Better In-Context Models.

Yaroslav Aksenov, Nikita Balagansky, Sofia Maria Lo Cicero Vaina, Boris Shaposhnikov, Alexey Gorbatovski, Daniil Gavrilov

Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), 2024

2023

Ahead-of-Time P-Tuning.

Daniil Gavrilov, Nikita Balagansky

CoRR, 2023

Democratized Diffusion Language Model.

Nikita Balagansky, Daniil Gavrilov

CoRR, 2023

2022

Linear Interpolation In Parameter Space is Good Enough for Fine-Tuned Language Models.

Mark Rofin, Nikita Balagansky, Daniil Gavrilov

CoRR, 2022

Classifiers are Better Experts for Controllable Text Generation.

Askhat Sitdikov, Nikita Balagansky, Daniil Gavrilov, Alexander Markov

CoRR, 2022

FastRPB: a Scalable Relative Positional Encoding for Long Sequence Tasks.

Maksim Zubkov, Daniil Gavrilov

CoRR, 2022

PALBERT: Teaching ALBERT to Ponder.

Nikita Balagansky, Daniil Gavrilov

Proceedings of the Advances in Neural Information Processing Systems 35: Annual Conference on Neural Information Processing Systems 2022, 2022

2021

Implicit Unlikelihood Training: Improving Neural Text Generation with Reinforcement Learning.

Evgeny Lagutin, Daniil Gavrilov, Pavel Kalaidin

Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume, 2021

2020

Weight Squeezing: Reparameterization for Compression and Fast Inference.

Artem V. Chumachenko, Daniil Gavrilov, Pavel Kalaidin

CoRR, 2020

2019

Self-attentive Model for Headline Generation.

Daniil Gavrilov, Pavel Kalaidin, Valentin Malykh

Proceedings of the Advances in Information Retrieval, 2019