SHARE: Social-Humanities AI for Research and Education

2026-04-13

AI Summary

This work addresses the absence of domain-specific pretrained language models for the social sciences and humanities (SSH), where general-purpose models often fail to capture disciplinary nuances and scholarly conventions. To bridge this gap, the authors introduce SHARE—a family of causal language models pretrained from scratch specifically for SSH—and accompany it with MIRROR, a non-generative interactive interface that enables users to critically review and engage with existing text without producing new content. The study also constructs a tailored Cloze evaluation benchmark for SSH, on which SHARE achieves performance comparable to Phi-4—a general model trained on over 100 times more tokens—demonstrating its efficiency and strong domain alignment. This approach establishes a novel generative AI interaction paradigm that adheres to SSH scholarly principles.

Technology Category

• Application Domains: Humanities & Computational Social Science
• Cognitive Modeling & Cognitive Systems: Social Cognition And Interaction
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Social Networks and Social Media: Generative AI / large language models and their impact on social systems
• Search and Retrieval-Augmented AI: Large language models for search
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

This intermediate technical report introduces the SHARE family of base models and the MIRROR user interface. The SHARE models are the first causal language models fully pretrained by and for the social sciences and humanities (SSH). Their performance in modelling SSH texts is close to that of general purpose models (Phi-4) which use 100 times more tokens, as shown by our custom SSH Cloze benchmark. The MIRROR user interface is designed for reviewing text inputs from the SSH disciplines while preserving critical engagement. By prototyping a generative AI interface that does not generate any text, we propose a way to harness the capabilities of the SHARE models without compromising the integrity of SSH principles and norms.

Problem

• social sciences and humanities
• generative AI
• critical engagement
• AI interface
• academic integrity

Innovation

• causal language models
• social sciences and humanities
• domain-specific pretraining
• non-generative interface
• critical engagement