conftrace_

shortcuts

3 papers

Explore in graph

Co-occurring keywords · papers mentioning both

group-wise losses (1) knowledge base complexity (1) role separation (1) token-wise cues (1) benchmark (1) pareto solution (1) position id (1) large language model (1) neural-symbolic reasoning (1) hypothesis space (1)

Papers

The Illusion of Role Separation: Hidden Shortcuts in LLM Role Learning (and How to Fix Them) ICML 2025 Analysis for Abductive Learning and Neural-Symbolic Reasoning Shortcuts ICML 2024 Improving Robustness to Multiple Spurious Correlations by Multi-Objective Optimization ICML 2024