Daniil Gavrilov
8 papers · 2021–2025 · 6 venues · across top CS/AI venues
Authors are matched by name, so researchers who share a name ("Yang Liu", "Wei Wang") appear as a single profile and their paper counts, venues and achievements are combined. How identity is handled.
Achievements
Jump to papers ↓
🌍
Conference Polyglot · Silver
🌉
Interdisciplinary Bridge · Bronze
🗺️
Taxonomy Completionist
(17)
⚡
Prolific Year · Productive
Top co-authors
Research topics
Artificial Intelligence
(7)
Core AI
(7)
Machine Learning
(6)
Deep Learning
(4)
Architectures
(3)
Natural Language Processing
(2)
Reinforcement Learning
(2)
Application Areas
(2)
Core Methods
(2)
Methods
(2)
Optimization & Theory · Machine Learning
(2)
Generation
(1)
Learning Types · Machine Learning
(1)
Learning Types · Deep Learning
(1)
Optimization & Theory · Deep Learning
(1)
Keywords
sparse autoencoder
(3)
language model
(2)
reinforcement learning
(2)
mathematical reasoning
(1)
direct preference optimization
(1)
in-context learning
(1)
chain-of-thought reasoning
(1)
text generation
(1)
language modeling
(1)
early exit
(1)
adaptive computation
(1)
state space model
(1)
mechanistic interpretability
(1)
sparsity constraint
(1)
steering vector
(1)
linear transformer
(1)
kernel function
(1)
inference speed
(1)
trust region optimization
(1)
cosine similarity
(1)
Papers
Analyze Feature Flow to Enhance Interpretation and Steering in Language Models
ICML 2025
Learn Your Reference Model for Real Good Alignment
ICLR 2025
Mechanistic Permutability: Match Features Across Layers
ICLR 2025
Steering LLM Reasoning Through Bias-Only Adaptation
EMNLP 2025
Train One Sparse Autoencoder Across Multiple Sparsity Budgets to Preserve Interpretability and Accuracy
EMNLP 2025
Linear Transformers with Learnable Kernel Functions are Better In-Context Models
ACL 2024
PALBERT: Teaching ALBERT to Ponder
NeurIPS 2022
Implicit Unlikelihood Training: Improving Neural Text Generation with Reinforcement Learning
EACL 2021