Entity · benchmark

Helpfulness Consistency

benchmarkactivehelpfulness-consistency-bc3f86ba·1 events·first seen May 22, 2026

Aliases: Helpfulness Consistency

Co-occurring entities

Sentiment Consistency Political Consistency Training (PCT)Reinforcement Learning

More like this (12)

Sentiment Consistency consistency training Consistency Training Can Entrench Misalignment Social Gaze Consistency Chain-of-Thought Self-Consistency operadic consistency Latent Consistency Models Dynamic-Probabilistic Consistency Gap HHH (Helpful, Harmless, Honest)Knowledge Assist Political Consistency Training (PCT)intervention-immediacy faithfulness

Recent events (1)

6arXiv · cs.AI·May 22, 2026·source ↗

Political Consistency Training: Reducing Covert Political Bias in LLMs via RL

Researchers identify a phenomenon called 'covert political bias' in LLMs, where models handle politically paired topics asymmetrically across 7 identified technique categories. They propose two metrics—Sentiment Consistency and Helpfulness Consistency—to measure this asymmetry. To address it, they introduce Political Consistency Training (PCT), an RL-based method with complementary training paradigms that reduces covert bias while preserving overall helpfulness and generalizing to held-out benchmarks.

Evaluation and Benchmarking AI Safety Research Sentiment Consistency Helpfulness Consistency Political Consistency Training (PCT)+2 more