Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

SciEx: Benchmarking Large Language Models on Scientific Exams with Human Expert Grading and Automatic Grading

Tu Anh Dinh, Carlos Mullov, Leonard Bärmann, Zhaolin Li, Danni Liu, Simon Reiß, Jueun Lee, Nathan Lerzer, Fabian Ternava, Jianfeng Gao, Tobias Röddiger, Alexander Waibel, Tamim Asfour, Michael Beigl, Rainer Stiefelhagen, Carsten Dachsbacher, Klemens Böhm, Jan Niehues

TL;DR

This paper proposes SciEx - a benchmark consisting of university computer science exam questions, to evaluate LLMs’ ability on solving scientific tasks, and proposes using LLM-as-a-judge to grade the LLM answers on SciEx.

Abstract

With the rapid development of Large Language Models (LLMs), it is crucial to have benchmarks which can evaluate the ability of LLMs on different domains. One common use of LLMs is performing tasks on scientific topics, such as writing algorithms, querying databases or giving mathematical proofs. Inspired by the way university students are evaluated on such tasks, in this paper, we propose SciEx - a benchmark consisting of university computer science exam questions, to evaluate LLMs ability on solving scientific tasks. SciEx is (1) multilingual, containing both English and German exams, and (2) multi-modal, containing questions that involve images, and (3) contains various types of freeform questions with different difficulty levels, due to the nature of university exams. We evaluate the performance of various state-of-the-art LLMs on our new benchmark. Since SciEx questions are freeform, it is not straightforward to evaluate LLM performance. Therefore, we provide human expert grading of the LLM outputs on SciEx. We show that the free-form exams in SciEx remain challenging for the current LLMs, where the best LLM only achieves 59.4\% exam grade on average. We also provide detailed comparisons between LLM performance and student performance on SciEx. To enable future evaluation of new LLMs, we propose using LLM-as-a-judge to grade the LLM answers on SciEx. Our experiments show that, although they do not perform perfectly on solving the exams, LLMs are decent as graders, achieving 0.948 Pearson correlation with expert grading.

SciEx: Benchmarking Large Language Models on Scientific Exams with Human Expert Grading and Automatic Grading

TL;DR

This paper proposes SciEx - a benchmark consisting of university computer science exam questions, to evaluate LLMs’ ability on solving scientific tasks, and proposes using LLM-as-a-judge to grade the LLM answers on SciEx.

Abstract

With the rapid development of Large Language Models (LLMs), it is crucial to have benchmarks which can evaluate the ability of LLMs on different domains. One common use of LLMs is performing tasks on scientific topics, such as writing algorithms, querying databases or giving mathematical proofs. Inspired by the way university students are evaluated on such tasks, in this paper, we propose SciEx - a benchmark consisting of university computer science exam questions, to evaluate LLMs ability on solving scientific tasks. SciEx is (1) multilingual, containing both English and German exams, and (2) multi-modal, containing questions that involve images, and (3) contains various types of freeform questions with different difficulty levels, due to the nature of university exams. We evaluate the performance of various state-of-the-art LLMs on our new benchmark. Since SciEx questions are freeform, it is not straightforward to evaluate LLM performance. Therefore, we provide human expert grading of the LLM outputs on SciEx. We show that the free-form exams in SciEx remain challenging for the current LLMs, where the best LLM only achieves 59.4\% exam grade on average. We also provide detailed comparisons between LLM performance and student performance on SciEx. To enable future evaluation of new LLMs, we propose using LLM-as-a-judge to grade the LLM answers on SciEx. Our experiments show that, although they do not perform perfectly on solving the exams, LLMs are decent as graders, achieving 0.948 Pearson correlation with expert grading.
Paper Structure (51 sections, 11 figures, 11 tables)

This paper contains 51 sections, 11 figures, 11 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. General-Purpose LLM Benchmarks
  4. Scientific LLM Benchmarks
  5. Freeform Answer Evaluation
  6. The SciEx Benchmark
  7. Univeristy Exams
  8. LLM-Generated Answers
  9. Expert Grading and Automatic Grading
  10. Data Creation
  11. Exam Collection
  12. Exam Formatting
  13. LLM-Generated Answers
  14. Expert Grading
  15. Automatic Grading
  16. ...and 36 more sections

Figures (11)

  • Figure 1: Question-level scores grouped by difficulty.
  • Figure 2: Difference between LLM scores and student scores, question-level, grouped by with/without images. Only Claude, GPT-4V and Llava can handle images.
  • Figure 3: Difference between LLM scores and student scores, question-level, grouped by languages.
  • Figure 4: Exam question before and after being converted to JSON format.
  • Figure 5: Answer generation prompt in English.
  • ...and 6 more figures