Closing the Context Gap: Activation Alignment for Tabular In-Context Learning

arXiv:2610.06679 · cs.LG, cs.AI · Submitted 2026-10-05 · Read on arXiv

cs.LG, cs.AI

Submitted: 2026-10-05

Updated: 2026-10-05

Code: https://github.com/yoel-zeldes/tabalign

Terminology

• Alignment — 1× in this paper · explained in 1 episode
• In-Context Learning — 1× in this paper · explained in 4 episodes

Sources

• Eliciting Latent Predictions from Transformers with the Tuned Lens
• Distilling the Knowledge in a Neural Network
• TACTICL: Task-Aware Compression of Tabular ICL Models
• FitNets: Hints for Thin Deep Nets
• Contrastive Representation Distillation
• Function Vectors in Large Language Models

Related papers

• Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
• AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
• Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
• Convergence issues in Relational Concept Analysis based on AOC-posets
• Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
• Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks