Design Principles and Guidelines for LLM Observability: Insights from Developers
Abstract
The growing number of large model development tools has increased the demand for transparency and accountability. We present four developer-centric design principles for large language model (LLM) observability across the entire LLMOps lifecycle. The research focuses on improving the developer experience by integrating design principles that address the unique needs of developers at various stages of LLM development. A mixed-methods methodology was implemented consisting of three focus groups (n=10 per group) with developers across different proficiency levels, complemented by a small-scale designer rating survey. The study highlights key design principles—Design for Awareness, Monitoring, Intervention, and Operability—and their impact on developer experience factors like mental models, efficiency, visualization, and understandability. The findings reveal significant correlations between these principles and developer needs, with a particular trade-off between efficiency and mental model clarity. The study provides actionable insights for designing effective observability tools and emphasizes the importance of balancing developer control, transparency, and ease of use.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
Public procurement of AI systems has relevance for several traditional human-centred domains, including community needs, transparency, AI governance, accountability, and policy.
“Although not commonly recognised as such, the public procurement of AI systems has relevance for several traditional human-centred domains, including community needs [84, 92, 103, 135], transparency [31, 145], AI governance [100, 107, 109, 169, 175], accountability [31, 131, 169, 175], and policy [75, 100, 103].”
From “It’s Just a Wild, Wild West”: Harnessing Public Procurement as an AI Governance Mechanism (2026) · cites this paper as [31]
Cites
Barredo Arrieta, A., Díaz-Rodríguez, N., Del Ser, J., Bennetot, A., Tabik, S., Barbado, A., Garcia, S., Gil-Lopez, S., Molina, D., Benjamins, R., et al. (2020) · Information Fusion · doi:10.1016/j.inffus.2019.12.012 · Cited by 10,423 · see citing works
Interpreting Black-Box Models: A Review on Explainable Artificial Intelligence
Hassija, V., Chamola, V., Mahapatra, A., Singal, A., Goel, D., Huang, K., Scardapane, S., Spinelli, I., Mahmud, M., & Hussain, A. (2024) · Cognitive Computation · doi:10.1007/s12559-023-10179-8 · Cited by 2,062 · see citing works
Re-examining Whether, Why, and How Human-AI Interaction Is Uniquely Difficult to Design
Yang, Q., Steinfeld, A., Rosé, C., & Zimmerman, J. (2020) · Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3313831.3376301 · Cited by 622 · cited 2× · see citing works
“These include widely applicable principles such as user interface design heuristics [16] and human-computer interaction strategies [17], as well as technology-specific guidelines like web usability standards [18] and human-AI interaction frameworks [19, 20].”
Nauta, M., Trienes, J., Pathak, S., Nguyen, E., Peters, M., Schmitt, Y., Schlötterer, J., van Keulen, M., & Seifert, C. (2023) · ACM Computing Surveys · doi:10.1145/3583558 · Cited by 590 · see citing works
Evaluating Explainable AI: Which Algorithmic Explanations Help Users Predict Model Behavior?
Hase, P., & Bansal, M. (2020) · Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics · doi:10.18653/v1/2020.acl-main.491 · Cited by 254 · see citing works
Flagged first, then by citations
Cited by
Sorted by citations
“It’s Just a Wild, Wild West”: Harnessing Public Procurement as an AI Governance Mechanism
Hudig, A. I., Kallina, E., & Singh, J. (2026) · Proceedings of the 2026 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3772318.3791968 · Cited by 3 · see citing works
- Authors
- Chen, X., Li, Y., & Wang, X.
- Affiliations
- Alibaba Group (China) (CN)
- Venue
- Proceedings of the Extended Abstracts of the CHI Conference on Human Factors in Computing Systems
- Publisher
- ACM
- Published
- 2025-04-25
- Pages
- 1-9
- OpenAlex
- W4409721409
- Cited by
- 7
- References
- 5
- Citations by year
- 2026: 5 · 2025: 2
- FWCI
- 8.69
- Type
- conference paper
- Language
- en
- Open access
- closed
- Locations
- CHI Conference Extended Abstracts on Human Factors in Computing Systems (CHI EA) (publishedVersion)
- Topics
- Physical Sciences › Computer Science › Computer Networks and Communications › Software System Performance and Reliability
- Physical Sciences › Computer Science › Information Systems › Software Engineering Techniques and Practices
- Physical Sciences › Computer Science › Information Systems › Software Engineering Research
- Keywords
- developer experience, design principles, transparency and accountability, mental models, visualization
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
