aiqbench analysis
This research introduces AttriWiki, a self-supervised pipeline that uses linear probes on hidden representations to distinguish between faithfulness and factuality errors in large language models. By accurately identifying the dominant knowledge source behind model outputs, this method significantly improves error detection rates without requiring retraining on conflicting datasets. The findings highlight a critical pathway for mitigating hallucinations by separating internal knowledge failures from context misuse.
Probing for Knowledge Attribution in Large Language Models
arXiv:2602.22787v3 Announce Type: replace-cross Abstract: Large language model (LLM) hallucinations, meaning fluent but factually incorrect generations, fall into two types: faithfulness violations, where the model misuses provided context, and factuality violations, where answers reflect errors in internal knowledge. Proper mitigation depends on knowing which source drives each answer. We study contributive attribution, i.e. the classification of the dominant knowledge source behind each output, and show that a simple linear probe trained on hidden representations can reliably identify it. We introduce AttriWiki, a self-supervised pipeline that automatically generates labelled training data by prompting models to recall withheld entities from memory or read them from context without relying on knowledge conflicts. Probes trained on AttriWiki achieve up to 0.96 Macro-$F_1$ on Llama-3.1-8B, Mistral-7B, and Qwen-7B, transfer to SQuAD and WebQuestions with 0.94-0.99 Macro-$F_1$, and generalise zero-shot to Tighidet et al. (2024)'s benchmark, outperforming their probe on conflicting settings without retraining. Furthermore, attribution mismatches raise error rates by up to 70%, though correct attribution does not guarantee correct answers, pointing to the need for broader detection frameworks.
Abstract:Large language model (LLM) hallucinations, meaning fluent but factually incorrect generations, fall into two types: faithfulness violations, where the model misuses provided context, and factuality violations, where answers reflect errors in internal knowledge. Proper mitigation depends on knowing which source drives each answer. We study contributive attribution, i.e. the classification of the dominant knowledge source behind each output, and show that a simple linear probe trained on hidden representations can reliably identify it. We introduce AttriWiki, a self-supervised pipeline that automatically generates labelled training data by prompting models to recall withheld entities from memory or read them from context without relying on knowledge conflicts. Probes trained on AttriWiki achieve up to 0.96 Macro-$F_1$ on Llama-3.1-8B, Mistral-7B, and Qwen-7B, transfer to SQuAD and WebQuestions with 0.94-0.99 Macro-$F_1$, and generalise zero-shot to Tighidet et al. (2024)'s benchmark, outperforming their probe on conflicting settings without retraining. Furthermore, attribution mismatches raise error rates by up to 70%, though correct attribution does not guarantee correct answers, pointing to the need for broader detection frameworks.
Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
arXiv:2602.22787 [cs.CL]
(or arXiv:2602.22787v3 [cs.CL] for this version)
https://doi.org/10.48550/arXiv.2602.22787 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 265 words
Related coverage
J++ Lens: Jacobian Filtering Enables More Faithful Workspace Lenses
LessWrong · Oct 8, 2026
Concrete Evaluations to Investigate the OpenAI Model That Hacked Hugging Face
AI Alignment Forum · Aug 3, 2026
[AINews] not much happened today
Latent Space · Sep 18, 2026
The Anatomy of a Truth Direction: Knowledge-Dependent Dimensionality, a Relational Law, and a Convergent Category Geometry in Small Language Models
arXiv · Jul 21, 2026