---
title: "Iterative Dual-Model Alignment for Story Evaluation — Qin 2026"
description: "Automatic story judges usually give verdicts without refining their reasoning. 22 references."
authors: ["Bruce Qin", "Dan Goldwasser"]
year: 2026
venue: "Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)"
source: "https://aclanthology.org/2026.acl-long.648/"
doi: "10.18653/v1/2026.acl-long.648"
licence: "CC BY 4.0 (ACL Anthology)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/qin-2026-iterative-dual-model-alignment-story/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/qin-2026-iterative-dual-model-alignment-story/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2026](https://md.research.elamur.ai/free/en/papers/2026/page/37/) / Iterative Dual-Model Alignment for Story Evaluation

# Iterative Dual-Model Alignment for Story Evaluation

**Authors:** Bruce Qin and [Dan Goldwasser](https://md.research.elamur.ai/free/en/authors/dan-goldwasser/) · **Year:** 2026 · **Venue:** [Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)](https://md.research.elamur.ai/free/en/venues/acl/) ([ACL 2026](https://md.research.elamur.ai/free/en/venues/acl/2026/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2026.acl-long.648/) · [PDF](https://aclanthology.org/2026.acl-long.648.pdf) · [DOI 10.18653/v1/2026.acl-long.648](https://doi.org/10.18653/v1/2026.acl-long.648) · [OpenAlex](https://openalex.org/W7166802488)

**Citations:** 0 citations (OpenAlex, 2026-10-14)

## Summary
Automatic story judges usually give verdicts without refining their reasoning. Critique generators, in contrast, produce feedback that is not tied to any consistent preference model. Qin and Goldwasser connect the two with an Iterative Alpha–Beta Learning framework made of two 8B models. α is a pairwise classifier that picks the more engaging of two prompt–story pairs. β writes rubric-structured comparisons of the two stories without naming a winner.

In each round, α scores β's candidate explanations by how much they raise the gold-label probability compared with no explanation. These rankings drive DPO updates for β. The top explanation is then used as auxiliary supervision when α is retrained with a KL-style cross-entropy loss.

On HANNA story pairs with human engagement labels, the joint system outperforms competitive baselines that rely on one model alone, with gains in accuracy and in the quality of its explanations across rounds. Its explanations help most on uncertain cases, are favored by experts, and lead to better story revisions. The work shows that explanations can be trained as working parts of an evaluator, using affordable mid-sized models instead of commercial APIs.

## Key points
- α is a causal LM used as a binary classifier. It reads the logits at the position after "Answer:" and applies a softmax over the A/B label tokens.
- Explanations are ranked by confidence uplift, Δ(E) = pθ(y|E) − pθ(y|∅).
- Experiments use only the HANNA dataset. The authors note its stories are short and sometimes repetitive or incoherent, which limits generalizability.
- In human evaluation, six NLP PhD annotators preferred trained β explanations 60.6% of the time, versus 39.4% for an untrained Llama-3.1-8B-Instruct baseline.
- Qualitatively, aligned β grounds its claims in specific textual cues, while the base model tends toward generic commentary. Code, models, data and prompts are promised on GitHub.

*Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.*

## Institutions

- [Purdue University West Lafayette](https://md.research.elamur.ai/free/en/universities/purdue-university-west-lafayette/) — University, West Lafayette, United States · 26 works · 2,076 citations

📚 Shelves (5): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 18,614 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 17,965 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 17,450 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 17,296 works · [👁️ Computer Vision and Pattern Recognition](https://md.research.elamur.ai/free/en/shelves/computer-science/computer-vision-and-pattern-recognition/) · 3,486 works

🧭 Topics (1): [📏 Evaluation & benchmarks](https://md.research.elamur.ai/free/en/topics/evaluation-benchmarks/) · 1,341 works

## Cite

Cite as: Bruce Qin and Dan Goldwasser. 2026. Iterative Dual-Model Alignment for Story Evaluation. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers).

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@inproceedings{qin2026iterative,
  title = {{Iterative Dual-Model Alignment for Story Evaluation}},
  author = {Bruce Qin and Dan Goldwasser},
  booktitle = {Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2026},
  doi = {10.18653/v1/2026.acl-long.648},
  url = {https://doi.org/10.18653/v1/2026.acl-long.648}
}
```

- [← previous iTAG: Inverse Design for Natural Text Generation with Accurate Causal Graph Annotations](https://md.research.elamur.ai/free/en/papers/wang-2026-itag-inverse-design-natural-text/)
- [next → Iterative Knowledge Graph Refinement and Integration for Medical Question Answering](https://md.research.elamur.ai/free/en/papers/zou-2026-iterative-knowledge-graph-refinement-integration/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/37/)

## References

22 references, 8 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Cyril Chhun, Pierre Colombo, Chloé Clavel, and Fabian M. Suchanek. 2022. Of human criteria and automatic metrics: A benchmark of the evaluation of story generation. In Proceedings of COLING. [find](https://scholar.google.com/scholar?q=Cyril+Chhun%2C+Pierre+Colombo%2C+Chlo%C3%A9+Clavel%2C+and+Fabian+M%2E+Suchanek%2E+2022%2E+Of+human+criteria+and+automatic+metrics%3A+A+benchmark+of+the+evaluation+of+story+generation%2E+In+Proceedings+of+COLING%2E)
2. Cyril Chhun, Fabian M Suchanek, and Chloé Clavel. 2024. [Do language models enjoy their own stories? prompting large language models for automatic story evaluation](https://md.research.elamur.ai/free/en/papers/chhun-2024-do-language-models-enjoy-their/). Transactions of the Association for Computational Linguistics, 12:1122–1142.
3. Paramveer S Dhillon, Somayeh Molaei, Jiaqi Li, Maximilian Golub, Shaochun Zheng, and Lionel Peter Robert. 2024. Shaping human-ai collaboration: Varied scaffolding levels in co-writing with language models. In Proceedings of the 2024 CHI Conference on Human Factors in Computing Systems, pages 1– 18. [find](https://scholar.google.com/scholar?q=Paramveer+S+Dhillon%2C+Somayeh+Molaei%2C+Jiaqi+Li%2C+Maximilian+Golub%2C+Shaochun+Zheng%2C+and+Lionel+Peter+Robert%2E+2024%2E+Shaping+human%2Dai+collaboration%3A+Varied+scaffolding+levels+in+co%2Dwriting+with+language+mo)
4. Daniel Fein, Sebastian Russo, Violet Xiang, Kabir Jolly, Rafael Rafailov, and Nick Haber. 2025. Litbench: A benchmark and dataset for reliable evaluation of creative writing. arXiv preprint arXiv:2507.00769. [find](https://scholar.google.com/scholar?q=Daniel+Fein%2C+Sebastian+Russo%2C+Violet+Xiang%2C+Kabir+Jolly%2C+Rafael+Rafailov%2C+and+Nick+Haber%2E+2025%2E+Litbench%3A+A+benchmark+and+dataset+for+reliable+evaluation+of+creative+writing%2E+arXiv+preprint+arXiv%3A2507)
5. Katy Ilonka Gero, Tao Long, and Lydia B. Chilton. 2023. Social dynamics of AI support in creative writing. In Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems, New York, NY, USA. Association for Computing Machinery. [find](https://scholar.google.com/scholar?q=Katy+Ilonka+Gero%2C+Tao+Long%2C+and+Lydia+B%2E+Chilton%2E+2023%2E+Social+dynamics+of+AI+support+in+creative+writing%2E+In+Proceedings+of+the+2023+CHI+Conference+on+Human+Factors+in+Computing+Systems%2C+New+York%2C+NY)
6. Jian Guan and Minlie Huang. 2020. Union: An unreferenced metric for evaluating open-ended story generation. arXiv preprint arXiv:2009.07602. [find](https://scholar.google.com/scholar?q=Jian+Guan+and+Minlie+Huang%2E+2020%2E+Union%3A+An+unreferenced+metric+for+evaluating+open%2Dended+story+generation%2E+arXiv+preprint+arXiv%3A2009%2E07602%2E)
7. Jian Guan, Zhexin Zhang, Zhuoer Feng, Zitao Liu, Wenbiao Ding, Xiaoxi Mao, Changjie Fan, and Minlie Huang. 2021. OpenMEVA: A benchmark for evaluating open-ended story generation metrics. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th Internati… [find](https://scholar.google.com/scholar?q=Jian+Guan%2C+Zhexin+Zhang%2C+Zhuoer+Feng%2C+Zitao+Liu%2C+Wenbiao+Ding%2C+Xiaoxi+Mao%2C+Changjie+Fan%2C+and+Minlie+Huang%2E+2021%2E+OpenMEVA%3A+A+benchmark+for+evaluating+open%2Dended+story+generation+metrics%2E+In+Proceeding)
8. Mina Lee, Percy Liang, and Qian Yang. 2022. [Coauthor: Designing a human–ai collaborative writing dataset for exploring language model capabilities](https://md.research.elamur.ai/free/en/papers/lee-2022-coauthor-designing-human-ai-collaborative/). In Proceedings of the 2022 CHI Conference on Human Factors in Computing Systems. ACM.
9. Chin-Yew Lin. 2004. [Rouge: A package for automatic evaluation of summaries](https://md.research.elamur.ai/free/en/papers/lin-2004-rouge-package-automatic-evaluation-summaries/). In Text summarization branches out: Proceedings of the ACL-04 workshop, pages 74–81.
10. Weize Liu, Guocong Li, Kai Zhang, Bang Du, Qiyuan Chen, Xuming Hu, Hongxia Xu, Jintai Chen, and Jian Wu. 2024. Mind’s mirror: Distilling self-evaluation capability and comprehensive thinking from large language models. In Proceedings of the 2024 Conference of the North American Chapter of the Associ… [find](https://scholar.google.com/scholar?q=Weize+Liu%2C+Guocong+Li%2C+Kai+Zhang%2C+Bang+Du%2C+Qiyuan+Chen%2C+Xuming+Hu%2C+Hongxia+Xu%2C+Jintai+Chen%2C+and+Jian+Wu%2E+2024%2E+Mind%E2%80%99s+mirror%3A+Distilling+self%2Devaluation+capability+and+comprehensive+thinking+from+larg)
11. Yang Liu, Dan Iter, Yichong Xu, Shuohang Wang, Ruochen Xu, and Chenguang Zhu. 2023. [G-eval: NLG evaluation using gpt-4 with better human alignment](https://md.research.elamur.ai/free/en/papers/liu-2023-g-eval-nlg-evaluation-gpt/). In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 2511–2522, Singapore. Association for Computational Lin…
12. Ryan Lowe, Michael Noseworthy, Iulian Vlad Serban, Nicolas Angelard-Gontier, Yoshua Bengio, and Joelle Pineau. 2017. [Towards an automatic turing test: Learning to evaluate dialogue responses](https://md.research.elamur.ai/free/en/papers/lowe-2017-towards-automatic-turing-test-learning/). In Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long P…
13. Piotr Mirowski, Kory W. Mathewson, Jaylen Pittman, and Richard Evans. 2022. Co-writing screenplays and theatre scripts with language models: An evaluation by industry professionals. In Proceedings of the Workshop on Machine Learning for Creativity and Design. [find](https://scholar.google.com/scholar?q=Piotr+Mirowski%2C+Kory+W%2E+Mathewson%2C+Jaylen+Pittman%2C+and+Richard+Evans%2E+2022%2E+Co%2Dwriting+screenplays+and+theatre+scripts+with+language+models%3A+An+evaluation+by+industry+professionals%2E+In+Proceedings+of+)
14. Kishore Papineni, Salim Roukos, Todd Ward, and Wei-Jing Zhu. 2002. [Bleu: a method for automatic evaluation of machine translation](https://md.research.elamur.ai/free/en/papers/papineni-2002-bleu-method-automatic-evaluation-machine/). In Proceedings of the 40th annual meeting of the Association for Computational Linguistics, pages 311–318. Association for Computational Linguistics.
15. Hannah Rashkin, Elizabeth Clark, Fantine Huot, and Mirella Lapata. 2025. Help me write a story: Evaluating LLMs’ ability to generate writing feedback. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 25827–25847, Vienna, Austri… [find](https://scholar.google.com/scholar?q=Hannah+Rashkin%2C+Elizabeth+Clark%2C+Fantine+Huot%2C+and+Mirella+Lapata%2E+2025%2E+Help+me+write+a+story%3A+Evaluating+LLMs%E2%80%99+ability+to+generate+writing+feedback%2E+In+Proceedings+of+the+63rd+Annual+Meeting+of+the+)
16. Ricardo Rei, Ana Farinha, Alon Lavie, and Andre FT Martins. 2020. [Comet: A neural framework for mt evaluation](https://md.research.elamur.ai/free/en/papers/rei-2020-comet-neural-framework-mt-evaluation/). In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP).
17. Thibault Sellam, Dipanjan Das, and Ankur Parikh. 2020. [Bleurt: Learning robust metrics for text generation](https://md.research.elamur.ai/free/en/papers/sellam-2020-bleurt-learning-robust-metrics-text/). In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics (ACL).
18. Chongyang Tao, Wei Wu, Can Xu, Yansong Zhang, Dongyan Zhao, and Rui Yan. 2018. Ruber: An unsupervised method for automatic evaluation of open-domain dialog systems. In Proceedings of the AAAI Conference on Artificial Intelligence. [find](https://scholar.google.com/scholar?q=Chongyang+Tao%2C+Wei+Wu%2C+Can+Xu%2C+Yansong+Zhang%2C+Dongyan+Zhao%2C+and+Rui+Yan%2E+2018%2E+Ruber%3A+An+unsupervised+method+for+automatic+evaluation+of+open%2Ddomain+dialog+systems%2E+In+Proceedings+of+the+AAAI+Conferen)
19. Maartje ter Hoeve, Julia Kiseleva, Maarten de Rijke, and 1 others. 2020. What makes a good summary? reconsidering the focus of automatic summarization. arXiv preprint arXiv:2012.07619. — [arXiv:2012.07619](https://arxiv.org/abs/2012.07619) [find](https://scholar.google.com/scholar?q=Maartje+ter+Hoeve%2C+Julia+Kiseleva%2C+Maarten+de+Rijke%2C+and+1+others%2E+2020%2E+What+makes+a+good+summary%3F+reconsidering+the+focus+of+automatic+summarization%2E+arXiv+preprint+arXiv%3A2012%2E07619%2E+%E2%80%94+%5BarXiv%3A2012%2E0)
20. Catherine Yeh, Gonzalo Ramos, Rachel Ng, Andy Huntington, and Richard Banks. 2024. Ghostwriter: Augmenting collaborative human-ai writing experiences through personalization and agency. arXiv preprint arXiv:2402.08855. — [arXiv:2402.08855](https://arxiv.org/abs/2402.08855) [find](https://scholar.google.com/scholar?q=Catherine+Yeh%2C+Gonzalo+Ramos%2C+Rachel+Ng%2C+Andy+Huntington%2C+and+Richard+Banks%2E+2024%2E+Ghostwriter%3A+Augmenting+collaborative+human%2Dai+writing+experiences+through+personalization+and+agency%2E+arXiv+preprint)
21. Ann Yuan, Andy Coenen, Emily Reif, and Daphne Ippolito. 2022. Wordcraft: story writing with large language models. In Proceedings of the 27th International Conference on Intelligent User Interfaces, pages 841–852. [find](https://scholar.google.com/scholar?q=Ann+Yuan%2C+Andy+Coenen%2C+Emily+Reif%2C+and+Daphne+Ippolito%2E+2022%2E+Wordcraft%3A+story+writing+with+large+language+models%2E+In+Proceedings+of+the+27th+International+Conference+on+Intelligent+User+Interfaces%2C+p)
22. Tianyi Zhang, Varsha Kishore, Felix Wu, Kilian Q Weinberger, and Yoav Artzi. 2019. Bertscore: Evaluating text generation with bert. arXiv preprint arXiv:1904.09675. [find](https://scholar.google.com/scholar?q=Tianyi+Zhang%2C+Varsha+Kishore%2C+Felix+Wu%2C+Kilian+Q+Weinberger%2C+and+Yoav+Artzi%2E+2019%2E+Bertscore%3A+Evaluating+text+generation+with+bert%2E+arXiv+preprint+arXiv%3A1904%2E09675%2E)

*Updated 2 November 2026*