Research Publications
Research Publication
How not to Lie with a Benchmark: Rearranging NLP Leaderboards.
arXiv (Cornell University)Dec 2, 20211 OpenAlex citationMatched to OpenAlex
Publication summary
Matched to OpenAlexNo abstract is available yet. The publication is shown because it is linked to an approved public researcher profile.
Topic ModelingNatural Language Processing TechniquesDomain Adaptation and Few-Shot LearningNLP benchmarksgeometric meanharmonic meanarithmetic meanmodel comparison
Authors
Related publications
Connected through authors and topics already present in the Research Hub.
Beyond Attention: Module-Level Refusal Directions in Mamba-Based Hybrid Language ModelsLecture notes in computer science · 2026 · 0 OpenAlex citationsSame author: Valentin MalykhA Reproducible Microscopy Acquisition Pipeline and Microscopy-Aware Attention for Classification, Detection, and Segmentation of MicroorganismsMachine Learning and Knowledge Extraction · 2026 · 0 OpenAlex citationsSame author: Valentin MalykhHybrid Non-local Blocks for Streptococci Segmentation in Microscopic ImagesLecture notes in networks and systems · 2026 · 0 OpenAlex citationsSame author: Valentin MalykhFlow-Guided Neural Pruning: Signal-Flow Framework for Multi-Architecture Model CompressionMachine Learning and Knowledge Extraction · 2026 · 0 OpenAlex citationsSame author: Valentin Malykh