UniReps @ NeurIPS 2025
82 papers
Top Research Directions
Research directions most represented by UniReps @ NeurIPS 2025 papers.
Top Authors
Impactful Papers
Papers from UniReps @ NeurIPS 2025 that shaped research clusters.
Foundational
Shared Parameter Subspaces and Cross-Task Linearity in Emergently Misaligned Behavior
foundational
Safety Degradation in Fine-Tuned Large L
Influential
On the Training Dynamics of Contrastive Learning with Imbalanced Feature Distributions: A Theoretica
influential
The Architectural Simplification Gap in
On the Impact of Topological Regularization on Geometrical and Topological Alignment in Autoencoders
influential
Mitigating Posterior Collapse in Latent
An Empirical Study of Task and Feature Correlations in the Reuse of Pre-trained Models
influential
The Real World Validation Gap of Synthet
Decoding Projections From Frozen Random Weights in Autoencoders: What Information Do They Encode?
influential
Greedy Layer-Wise Unsupervised Pre-Train
Shared Parameter Subspaces and Cross-Task Linearity in Emergently Misaligned Behavior
influential
Low-Rank Weight Decomposition for Effici
Look, Then Speak: Social Tokens for Grounding LLMs in Visual Interactions
influential
Aligning Visual Features with Language M
Papers
« prev Page 2/2