"Check My Work?": Measuring Sycophancy in a Simulated Educational Context

• URL: http://arxiv.org/abs/2506.10297v1
• Date: Thu, 12 Jun 2025 02:21:43 GMT
• Title: "Check My Work?": Measuring Sycophancy in a Simulated Educational Context
• Authors: Chuck Arvin,
• Abstract summary: This study examines how user-provided suggestions affect Large Language Models (LLMs) in a simulated educational context.<n>We show that response quality varies dramatically based on query framing.<n>Our results highlight the need to better understand the mechanism, and ways to mitigate, such bias in the educational context.
• Score: 0.0
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: This study examines how user-provided suggestions affect Large Language Models (LLMs) in a simulated educational context, where sycophancy poses significant risks. Testing five different LLMs from the OpenAI GPT-4o and GPT-4.1 model classes across five experimental conditions, we show that response quality varies dramatically based on query framing. In cases where the student mentions an incorrect answer, the LLM correctness can degrade by as much as 15 percentage points, while mentioning the correct answer boosts accuracy by the same margin. Our results also show that this bias is stronger in smaller models, with an effect of up to 30% for the GPT-4.1-nano model, versus 8% for the GPT-4o model. Our analysis of how often LLMs "flip" their answer, and an investigation into token level probabilities, confirm that the models are generally changing their answers to answer choices mentioned by students in line with the sycophancy hypothesis. This sycophantic behavior has important implications for educational equity, as LLMs may accelerate learning for knowledgeable students while the same tools may reinforce misunderstanding for less knowledgeable students. Our results highlight the need to better understand the mechanism, and ways to mitigate, such bias in the educational context.

Related papers

• Can LLMs Reliably Simulate Real Students' Abilities in Mathematics and Reading Comprehension? [8.558834738072363]
  Large Language Models (LLMs) are increasingly used as proxy students in the development of Intelligent Tutoring Systems (ITSs)<n>We collect a dataset of 489 items from the National Assessment of Educational Progress (NAEP) covering mathematics and reading comprehension in grades 4, 8, and 12.<n>We apply an Item Response Theory (IRT) model to position 11 diverse and state-of-the-art LLMs on the same ability scale as real student populations.
  arXiv  Detail & Related papers  (2025-07-11T00:36:57Z)
• Unleashing the Reasoning Potential of Pre-trained LLMs by Critique Fine-Tuning on One Problem [53.3188041952701]
  We show that Critique Fine-Tuning (CFT) on only one problem can effectively unleash the reasoning potential of LLMs.<n>With just 5 GPU hours of training, Qwen-Math-7B-CFT show an average improvement of 15% on six math benchmarks and 16% on three logic reasoning benchmarks.<n>Results are comparable to or even surpass the results from RL with 20x less compute.
  arXiv  Detail & Related papers  (2025-06-03T18:35:52Z)
• Can Large Language Models Match Tutoring System Adaptivity? A Benchmarking Study [0.0]
  Large Language Models (LLMs) hold promise as dynamic instructional aids.<n>Yet, it remains unclear whether LLMs can replicate the adaptivity of intelligent tutoring systems (ITS)
  arXiv  Detail & Related papers  (2025-04-07T23:57:32Z)
• LLM Robustness Against Misinformation in Biomedical Question Answering [50.98256373698759]
  The retrieval-augmented generation (RAG) approach is used to reduce the confabulation of large language models (LLMs) for question answering. We evaluate the effectiveness and robustness of four LLMs against misinformation in answering biomedical questions.
  arXiv  Detail & Related papers  (2024-10-27T16:23:26Z)
• Learning to Love Edge Cases in Formative Math Assessment: Using the AMMORE Dataset and Chain-of-Thought Prompting to Improve Grading Accuracy [0.0]
  This paper introduces AMMORE, a new dataset of 53,000 math open-response question-answer pairs from Rori. We conduct two experiments to evaluate the use of large language models (LLM) for grading challenging student answers.
  arXiv  Detail & Related papers  (2024-09-26T14:51:40Z)
• See What LLMs Cannot Answer: A Self-Challenge Framework for Uncovering LLM Weaknesses [51.975495361024606]
  We propose a Self-Challenge evaluation framework with human-in-the-loop. Starting from seed instances that GPT-4 fails to answer, we prompt GPT-4 to summarize error patterns that can be used to generate new instances. We then build a benchmark, SC-G4, consisting of 1,835 instances generated by GPT-4 using these patterns, with human-annotated gold responses.
  arXiv  Detail & Related papers  (2024-08-16T19:01:52Z)
• Perception of Knowledge Boundary for Large Language Models through Semi-open-ended Question Answering [67.94354589215637]
  Large Language Models (LLMs) are widely used for knowledge-seeking yet suffer from hallucinations. In this paper, we perceive the LLMs' knowledge boundary (KB) with semi-open-ended questions (SoeQ) We find that GPT-4 performs poorly on SoeQ and is often unaware of its KB. Our auxiliary model, LLaMA-2-13B, is effective in discovering more ambiguous answers.
  arXiv  Detail & Related papers  (2024-05-23T10:00:14Z)
• Language Models can Evaluate Themselves via Probability Discrepancy [38.54454263880133]
  We propose a new self-evaluation method ProbDiff for assessing the efficacy of various Large Language Models (LLMs) It uniquely utilizes the LLMs being tested to compute the probability discrepancy between the initial response and its revised versions. Our findings reveal that ProbDiff achieves results on par with those obtained from evaluations based on GPT-4.
  arXiv  Detail & Related papers  (2024-05-17T03:50:28Z)
• How Easy is It to Fool Your Multimodal LLMs? An Empirical Analysis on Deceptive Prompts [54.07541591018305]
  We present MAD-Bench, a benchmark that contains 1000 test samples divided into 5 categories, such as non-existent objects, count of objects, and spatial relationship. We provide a comprehensive analysis of popular MLLMs, ranging from GPT-4v, Reka, Gemini-Pro, to open-sourced models, such as LLaVA-NeXT and MiniCPM-Llama3. While GPT-4o achieves 82.82% accuracy on MAD-Bench, the accuracy of any other model in our experiments ranges from 9% to 50%.
  arXiv  Detail & Related papers  (2024-02-20T18:31:27Z)
• Are You Sure? Challenging LLMs Leads to Performance Drops in The FlipFlop Experiment [82.60594940370919]
  We propose the FlipFlop experiment to study the multi-turn behavior of Large Language Models (LLMs) We show that models flip their answers on average 46% of the time and that all models see a deterioration of accuracy between their first and final prediction, with an average drop of 17% (the FlipFlop effect) We conduct finetuning experiments on an open-source LLM and find that finetuning on synthetically created data can mitigate - reducing performance deterioration by 60% - but not resolve sycophantic behavior entirely.
  arXiv  Detail & Related papers  (2023-11-14T23:40:22Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.