RETUYT-INCO at BEA 2026 Shared Task 2: Meta-prompting in Rubric-based Scoring for German

arXiv cs.CL Papers
shared-task automatic-grading meta-prompting german-nlp educational-ai

Summary

This paper details the RETUYT-INCO team's participation in the BEA 2026 Shared Task 2, introducing a meta-prompting approach for rubric-based scoring of German short answers.

arXiv:2605.11242v1 Announce Type: new Abstract: In this paper, we present the RETUYT-INCO participation at the BEA 2026 shared task "Rubric-based Short Answer Scoring for German". Our team participated in track 1 (Unseen answers three-way), track 3 (Unseen answers two-way) and track 4 (Unseen questions two-way). Since these tracks required scoring short student answers using specific rubrics, we looked for ways to handle the changing nature of the task. We created a method called Meta-prompting. In this approach, an LLM creates a custom prompt based on examples from the Train set. This prompt is then used to grade new student answers. Along with this method, we also describe other approaches we used, such as classic machine learning, fine-tuning open-source LLMs, and different prompting techniques. According to the official results, our team placed 6th out of 8 participants in Track 1 with a QWK of 0.729. In Track 3, we secured 4th place out of 9 with a QWK of 0.674, and we also placed 4th out of 8 in Track 4 with a QWK of 0.49.
Original Article
View Cached Full Text

Cached at: 05/13/26, 06:09 AM 

# RETUYT-INCO at BEA 2026 Shared Task 2: Meta-prompting in Rubric-based Scoring for German
Source: [https://arxiv.org/abs/2605.11242](https://arxiv.org/abs/2605.11242)
[View PDF](https://arxiv.org/pdf/2605.11242)

> Abstract:In this paper, we present the RETUYT\-INCO participation at the BEA 2026 shared task "Rubric\-based Short Answer Scoring for German"\. Our team participated in track 1 \(Unseen answers three\-way\), track 3 \(Unseen answers two\-way\) and track 4 \(Unseen questions two\-way\)\. Since these tracks required scoring short student answers using specific rubrics, we looked for ways to handle the changing nature of the task\. We created a method called Meta\-prompting\. In this approach, an LLM creates a custom prompt based on examples from the Train set\. This prompt is then used to grade new student answers\. Along with this method, we also describe other approaches we used, such as classic machine learning, fine\-tuning open\-source LLMs, and different prompting techniques\. According to the official results, our team placed 6th out of 8 participants in Track 1 with a QWK of 0\.729\. In Track 3, we secured 4th place out of 9 with a QWK of 0\.674, and we also placed 4th out of 8 in Track 4 with a QWK of 0\.49\.

## Submission history

From: Santiago Góngora \[[view email](https://arxiv.org/show-email/3567939e/2605.11242)\] **\[v1\]**Mon, 11 May 2026 21:01:05 UTC \(61 KB\)

Similar Articles

RECAP: Regression Evaluation for Continual Adaptation of Prompts

arXiv cs.LG

Introduces RECAP, a benchmark for evaluating continual learning of prompts under evolving constraints in a proactive adaptation setting. Results show that existing prompt optimization methods fail in this setting, highlighting the need for new methods.

Prompt-Level Reward Specifications for Open-Ended Post-Training

arXiv cs.CL

This paper proposes a prompt-level reward specification framework that separates reward specification from computation, constructing reusable task-adaptive rubrics and executable constraint checkers offline to produce a hybrid reward for open-ended post-training without requiring human annotations or separate reward models.

Rubric-Aligned Disentangled Evaluation of Human Simultaneous Interpreting

arXiv cs.CL

This paper introduces a rubric-aligned method for evaluating human simultaneous interpreting at the segment level, using a LoRA-adapted COMET-KIWI model to disentangle dimensions like meaning transfer and delivery quality, achieving improved correlation with human ratings over baseline methods.