Nikita Balagansky
7 papers · 2022–2025 · 5 conferences · across top CS/AI conferences
Achievements
Jump to papers ↓+1 more ↓ Show less ↑
π Conference Polyglot (5) π Interdisciplinary Bridge πΊοΈ Taxonomy Completionist (12) π§ Keyword Pioneer π Cross-Pollinator (15)
β‘
Prolific Year
(5)
Conferences
EMNLP (2)
ICLR (2)
ACL (1)
ICML (1)
NIPS (1)
Top co-authors
Keywords
model compression
(1)
reinforcement learning
(1)
mathematical reasoning
(1)
in-context learning
(1)
chain-of-thought reasoning
(1)
language modeling
(1)
early exit
(1)
adaptive computation
(1)
state space model
(1)
language model
(1)
sparse autoencoder
(1)
sparsity constraint
(1)
steering vector
(1)
linear transformer
(1)
kernel function
(1)
inference speed
(1)
hierarchical training
(1)
interpretable feature
(1)
learnable kernel
(1)
bias-only adaptation
(1)
Papers
Train One Sparse Autoencoder Across Multiple Sparsity Budgets to Preserve Interpretability and Accuracy
EMNLP 2025
Mechanistic Permutability: Match Features Across Layers
ICLR 2025
Analyze Feature Flow to Enhance Interpretation and Steering in Language Models
ICML 2025
Steering LLM Reasoning Through Bias-Only Adaptation
EMNLP 2025
Learn Your Reference Model for Real Good Alignment
ICLR 2025
Linear Transformers with Learnable Kernel Functions are Better In-Context Models
ACL 2024
PALBERT: Teaching ALBERT to Ponder
NIPS 2022