BridgeAlign: Bridging Preference Alignment for Humanities and Social Sciences
Quick Answer
BridgeAlign introduces a novel preference-alignment pipeline for humanities and social sciences, enhancing Qwen3-8B's performance across 17 benchmarks.
Quick Take
It curates over 210k synthetic preference samples, achieving superior human-preference and knowledge-based capabilities without trade-offs, addressing the need for nuanced quality judgments in open-ended domains.
Key Points
- BridgeAlign consists of three phases: Seed Curation, Preference Data Synthesis, and Preference Optimization.
- The model aligns over 210k synthetic preference samples, setting new benchmarks in HSS tasks.
- It achieves the best average performance against 11 strong baselines across 17 benchmarks.
- No trade-off exists between human-preference and knowledge-based capabilities in the results.
- The approach addresses the unique challenges of quality judgments in humanities and social sciences.
Paper Resources
📖 Reader Mode
~2 min readAuthors:Ru Peng, Haokai Xu, Xijun Gu, Tianyu Zhao, Zhiting Fan, Yawen Zeng, Yihong Zhuang, Jinyang Zhang, Kexin Yang, Jian Wu, Hao Chen, Junyang Lin, Dayiheng Liu, Junbo Zhao
Abstract:While data synthesis for large language models (LLMs) is prevalent, it primarily targets domains with verifiable answers, overlooking open-ended humanities and social sciences (HSS), where nuanced quality judgments matter more than objective correctness. This makes preference alignment a natural paradigm for broad HSS tasks. Yet existing methods are either costly or not tailored to broad HSS disciplines. We thus propose BridgeAlign, among the first preference-alignment pipelines for broad HSS disciplines, with three phases: i) Seed Curation: curating HSS seed documents from web corpora via heuristic/LLM-based filtering and text refinement; ii) Preference Data Synthesis: generating preference triplets via persona-based instruction inversion with Q&A consistency checks; iii) Preference Optimization: moving beyond naive human-vs-model heuristics by first grounding preferences in HSS quality rubric, then generating transitional responses via controlled quality degradation to form near-boundary preference pairs for finer-grained quality discrimination. Aligning over 210k synthetic preference samples, BridgeAlign enables Qwen3-8B to achieve the best average across 17 benchmarks against 11 strong baselines; importantly, leading on both human-preference and knowledge-based capabilities at once, with no trade-off between them, as supported by extensive experiments and contextualized by existing theories.
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2607.27366 [cs.CL] |
| (or arXiv:2607.27366v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2607.27366 arXiv-issued DOI via DataCite |
Submission history
From: Ru Peng [view email]
[v1]
Wed, 29 Jul 2026 18:22:50 UTC (3,487 KB)
— Originally published at arxiv.org
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.CL
See more →TriAgent: Divergence-Aware Committees for Cost-Efficient Financial Sentiment Analysis
TriAgent introduces a cost-efficient multi-agent system for financial sentiment analysis, combining VADER, FinBERT, and Qwen2.5. It achieves an F1 score of ~0.87 with significant savings of $9.3M/year at a 10M-user scale compared to GPT-4o-mini, while also detecting hallucinations with an AUC of 0.90.