← Back to all news
arXiv Machine Learning

Unsupervised Identification and Removal of Spurious Correlations During Fine-Tuning

The paper presents GRASP, a method for unsupervised detection and removal of spurious correlations that arise during fine‑tuning of pretrained language models. By identifying latent factors from LoRA fine‑tune weights and projecting gradients to prevent new reliance on them, GRASP preserves genuine task signals while mitigating bias. Experiments on three tasks—misaligned code generation, medical advice, and political bias—show that GRASP reduces misalignment and drift more effectively than existing baselines while maintaining task performance.

Related stories

arXiv Computation and Language
Sep 18

The Neutral Mask: How Alignment Training Provides Shallow Alignment while Leaving Partisan Structure Intact in a Large Language Model

The paper investigates how alignment training, specifically reinforcement learning from human feedback (RLHF), affects the internal partisan structure of a large language model. Using a mechanistic case study on Llama 3.1 8B, the authors find that alignment training does not erase the model’s partisan geometry but compresses its variance, producing consistently balanced, non‑partisan outputs. Sparse autoencoder analysis and feature‑level steering experiments reveal that policy‑encoding features become inactive in the aligned model, indicating a causal disconnect rather than structural removal of partisan knowledge.

By Wendy K. Tam
llmsreinforcement-learningsafety
More like this →
arXiv AI
Sep 2

BiasGym: A Simple and Generalizable Framework for Analyzing and Removing Biases through Injection

BiasGym is a cost‑effective, generalizable framework that injects specific biases into large language models via token‑based fine‑tuning while keeping the model frozen. It then uses two debiasing methods—Scope and Steer—to identify and suppress or redirect the components responsible for biased behavior. The framework enables consistent bias elicitation, precise localization of bias associations, and targeted debiasing without harming downstream performance, and it has been shown to reduce real‑world stereotypes such as labeling Italians as reckless drivers.

By Sekh Mainul Islam, Nadav Borenstein, Siddhesh Milind Pawar, Haeun Yu, Arnav Arora, Isabelle Augenstein
llmsfine-tuningsafety
More like this →
arXiv Machine Learning
3d ago

ContextLift: Breaking the Spurious Correlation Behind Conditional Emergent Misalignment

ContextLift is a counterfactual augmentation technique that mitigates conditional emergent misalignment in language models. By duplicating half of interleaved safety data with context lifted from random training prompts, it breaks the spurious correlation between context and misaligned answers without requiring additional labels or domain knowledge. Experiments on Qwen2.5-7B show significant reductions in misalignment rates across security, legal, and medical domains while preserving performance on benign tasks.

By Florian Mai, David Kacz\'er
llmsfine-tuningsafety
More like this →
arXiv Computation and Language
Aug 27

Mitigating LLM biases toward spurious social contexts using direct preference optimization

The paper examines how large language models (LLMs) can be biased by irrelevant social contexts when evaluating teachers, using a large U.S. classroom transcript dataset. It shows that spurious contexts can shift model ratings by up to 1.48 points on a 7‑point scale and that standard mitigation methods like SFT and DPO are insufficient. The authors introduce Debiasing‑DPO, a method that combines contrastive reasoning‑augmented DPO with SFT, which reduces bias by 84% and improves predictive accuracy by 52% on Llama and Qwen Instruct models.

By Hyunji Nam, Dorottya Demszky
llmsreinforcement-learningfine-tuningsafety
More like this →