# Challenges in Trustworthy Human Evaluation of Chatbots

arXiv: https://arxiv.org/abs/2412.04363 · HTML: https://mlsanity.com/paper/2412.04363/

## Key facts

- Cited by 14 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 5 Dec 2024.

## Built on by (most cited first, 14 of 14)

- [EvalTree: Profiling Language Model Weaknesses via Hierarchical Capability Trees](https://mlsanity.com/paper/2503.08893/) (2025)
- [Prototypical Human-AI Collaboration Behaviors from LLM-Assisted Writing in the Wild](https://mlsanity.com/paper/2505.16023/) (2025)
- [Exploring and Mitigating Adversarial Manipulation of Voting-Based Leaderboards](https://mlsanity.com/paper/2501.07493/) (2025)
- [Improving Your Model Ranking on Chatbot Arena by Vote Rigging](https://mlsanity.com/paper/2501.17858/) (2025)
- [Dropping Just a Handful of Preferences Can Change Top Large Language Model Rankings](https://mlsanity.com/paper/2508.11847/) (2025)
- [Exploiting Leaderboards for Large-Scale Distribution of Malicious Models](https://mlsanity.com/paper/2507.08983/) (2025)
- [Curiosity by Design: An LLM-based Coding Assistant Asking Clarification Questions](https://mlsanity.com/paper/2507.21285/) (2025)
- [A Unified Perturbation Framework for Analyzing Leaderboard Stability and Manipulation](https://mlsanity.com/paper/2605.15761/) (2026)
- [Beyond Ordinal Preferences: Why Alignment Needs Cardinal Human Feedback](https://mlsanity.com/paper/2508.08486/) (2025)
- [Text-to-Image Models Leave Identifiable Signatures: Implications for Leaderboard Security](https://mlsanity.com/paper/2510.06525/) (2025)
- [Identifying Models Behind Text-to-Image Leaderboards](https://mlsanity.com/paper/2601.09647/) (2026)
- [InterPol: De-anonymizing LM Arena via Interpolated Preference Learning](https://mlsanity.com/paper/2603.15220/) (2026)
- [CV-Arena: An Open Benchmark for Instructional Computer Vision Problem Solving with Human-AI Collaborative Preferences](https://mlsanity.com/paper/2606.00931/) (2026)
- [Gaming Consensus: Coordinated Manipulation in Crowdsourced Fact-Checking](https://mlsanity.com/paper/2607.01824/) (2026)

## Abstract (arXiv)

Open community-driven platforms like Chatbot Arena that collect user preference data from site visitors have gained a reputation as one of the most trustworthy publicly available benchmarks for LLM performance. While now standard, it is tricky to implement effective guardrails to collect high-quality annotations from humans. In this paper, we demonstrate that three sources of bad annotations, both malicious and otherwise, can corrupt the reliability of open leaderboard rankings. In particular, we show that only 10% of poor quality votes by apathetic (site visitors not appropriately incentivized to give correct votes) or adversarial (bad actors seeking to inflate the ranking of a target model) annotators can change the rankings of models by up to 5 places on the leaderboard. Finally, we discuss open challenges in ensuring high-quality human annotations.