CopeNLU presents 7 papers at EMNLP 2026 on bias, alignment, explainability
IAugenstein · x · 2026-10-30
University of Copenhagen's CopeNLU group presented 7 main-conference papers, 3 workshop papers, and a workshop keynote at EMNLP 2026 in Budapest.
Main-conference work includes measuring epistemic diversity in LLMs, a communicative audit of LLM response framing, and disentangling cultural vs. personal alignment. Findings papers cover BiasGym (a generalizable framework for analyzing and removing biases via injection), human vs. LLM annotation in active learning for TikTok hate speech detection, and the impact of quantization on LLM self-explanations. Topics span explainability, bias detection, cross-cultural alignment, and limited-data learning.
More from Research
- AIR: Effective-Learning-Rate Scheduler Beats Cosine on 128B-Token MoE Pretraining — stochasticchasm · 2026-10-30
- NVIDIA researcher: train robot policies with privileged sim info, improving Pi 0.5 — wellecks · 2026-10-30
- CMU's Gym-Anything Auto-Builds 200+ Agent Environments from Any Software, Selected by GDP — wellecks · 2026-10-30
- CMU Auto-Builds 25 Robot Simulations with Agents; Frontier Model Solves 0 of 25 Test Tasks — wellecks · 2026-10-30
- Leveraging Simulator Privileged Information: Three Methods to Improve Robot Models — wellecks · 2026-10-30
- DeepMind's f_charton shares his take on the current state of AI and mathematics in Sorbonne talk — f_charton · 2026-10-30