BabyLMs: A Low‑Cost Sandbox to Study and Fix Bias in Language Models

Kari Jaaskelainen

15 Jan 2026 — 1 min read

TL;DR

Debiasing big language models is costly. This study shows compact “BabyLMs” can mimic how larger BERT-style models learn biases—so researchers can test ideas faster and cheaper.

BabyLMs (small, BERT-like models on tiny, editable corpora) track the same bias and performance patterns as standard BERTs.
Correlations hold across multiple debiasing strategies, both pre-training and post-hoc.
Using BabyLMs, the authors replicate past findings and reveal how gender imbalance and toxic text in training data drive bias.
Compute savings: from 500+ GPU-hours to under 30 for pre-training experiments.

Why it matters: A practical, compute-efficient sandbox to explore fairer training recipes—opening pre-training debiasing to more labs, students, and civic groups.

Paper by Filip Trhlik, Andrew Caines, and Paula Buttery (cs.CL/cs.AI). Read more: https://arxiv.org/abs/2601.09421v1

Paper: https://arxiv.org/abs/2601.09421v1

Register: https://www.AiFeta.com

AI NLP EthicalAI Bias MachineLearning BERT LLM Fairness Research

Meet GRACE: a moral governor for safer, more transparent AI

AI agents are getting powerful—so how do we make sure they do the right thing, not just the effective thing? Meet GRACE, a reason-based moral governor that keeps AI behavior aligned with human norms by separating moral reasoning from goal-driven decision-making. * Moral Module: uses deontic logic and explicit reasons

Why some fine-tuned LLMs miss phishing—and how to fix it

Not all fine-tuned LLMs spot phishing equally. A new study tests Llama 3.1 8B, Gemma 2 9B, and Mistral on high-stakes phishing detection—and uses SHAP and mechanistic interpretability to reveal why models do (or don’t) generalize. * Architecture × data diversity matters: Gemma 2 9B hits state-of-the-art performance (F1

AI that explains itself—by following the science

AI that explains itself—by following the science Black-box AI can be powerful, but it often can’t tell us why it made a decision. Concept Bottleneck Models (CBMs) try to fix that by predicting human-understandable concepts first, then the final answer. The catch: standard CBMs ignore domain-specific cause-and-effect and

Meet GenomAgent: A Team of AI Specialists for Smarter Genomics Q&A

TL;DR: Finding reliable answers in genomics is hard. GenomAgent turns one big AI into a coordinating team of specialists—and beats the current leader by 12% on a key benchmark. Genomic facts live across many databases. Standard chatbots struggle because they can’t flexibly query those sources. GeneGPT added