# TAXI: Evaluating Categorical Knowledge Editing for Language Models

arXiv: https://arxiv.org/abs/2404.15004 · HTML: https://mlsanity.com/paper/2404.15004/

## Key facts

- Cited by 7 papers in the map (data through 14 Oct 2026).
- Builds on 19 of its 30 references that are in the map.
- First posted on arXiv 23 Apr 2024.

## Builds on (most cited first, 19 of 19)

- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [Locating and Editing Factual Associations in GPT](https://mlsanity.com/paper/2202.05262/) (2022)
- [When Not to Trust Language Models: Investigating Effectiveness of Parametric and Non-Parametric Memories](https://mlsanity.com/paper/2212.10511/) (2022)
- [Mass-Editing Memory in a Transformer](https://mlsanity.com/paper/2210.07229/) (2022)
- [Editing Factual Knowledge in Language Models](https://mlsanity.com/paper/2104.08164/) (2021)
- [Dolma: an Open Corpus of Three Trillion Tokens for Language Model Pretraining Research](https://mlsanity.com/paper/2402.00159/) (2024)
- [Memory-Based Model Editing at Scale](https://mlsanity.com/paper/2206.06520/) (2022)
- [The Reversal Curse: LLMs trained on "A is B" fail to learn "B is A"](https://mlsanity.com/paper/2309.12288/) (2023)
- [MQuAKE: Assessing Knowledge Editing in Language Models via Multi-Hop Questions](https://mlsanity.com/paper/2305.14795/) (2023)
- [Can We Edit Factual Knowledge by In-Context Learning?](https://mlsanity.com/paper/2305.12740/) (2023)
- [Aging with GRACE: Lifelong Model Editing with Discrete Key-Value Adaptors](https://mlsanity.com/paper/2211.11031/) (2022)
- [Evaluating the Ripple Effects of Knowledge Editing in Language Models](https://mlsanity.com/paper/2307.12976/) (2023)
- [Transformer-Patcher: One Mistake worth One Neuron](https://mlsanity.com/paper/2301.09785/) (2023)
- [Knowledge Editing for Large Language Models: A Survey](https://mlsanity.com/paper/2310.16218/) (2023)
- [Infini-gram: Scaling Unbounded n-gram Language Models to a Trillion Tokens](https://mlsanity.com/paper/2401.17377/) (2024)
- [Detecting Edit Failures In Large Language Models: An Improved Specificity Benchmark](https://mlsanity.com/paper/2305.17553/) (2023)
- [A Survey on Knowledge Editing of Neural Networks](https://mlsanity.com/paper/2310.19704/) (2023)
- [Emptying the Ocean with a Spoon: Should We Edit Models?](https://mlsanity.com/paper/2310.11958/) (2023)
- [On the Robustness of Editing Large Language Models](https://mlsanity.com/paper/2402.05827/) (2024)

## Built on by (most cited first, 7 of 7)

- [Can Knowledge Editing Really Correct Hallucinations?](https://mlsanity.com/paper/2410.16251/) (2024)
- [Fundamental Problems With Model Editing: How Should Rational Belief Revision Work in LLMs?](https://mlsanity.com/paper/2406.19354/) (2024)
- [Can Editing LLMs Inject Harm?](https://mlsanity.com/paper/2407.20224/) (2024)
- [CodeUpdateArena: Benchmarking Knowledge Editing on API Updates](https://mlsanity.com/paper/2407.06249/) (2024)
- [Characterizing the Role of Similarity in the Property Inferences of Language Models](https://mlsanity.com/paper/2410.22590/) (2024)
- [CaKE: Circuit-aware Editing Enables Generalizable Knowledge Learners](https://mlsanity.com/paper/2503.16356/) (2025)
- [Composable Interventions for Language Models](https://mlsanity.com/paper/2407.06483/) (2024)

## Abstract (arXiv)

Humans rarely learn one fact in isolation. Instead, learning a new fact induces knowledge of other facts about the world. For example, in learning a korat is a type of cat, you also infer it is a mammal and has claws, ensuring your model of the world is consistent. Knowledge editing aims to inject new facts into language models to improve their factuality, but current benchmarks fail to evaluate consistency, which is critical to ensure efficient, accurate, and generalizable edits. We manually create TAXI, a new benchmark dataset specifically created to evaluate consistency in categorical knowledge edits. TAXI contains 11,120 multiple-choice queries for 976 edits spanning 41 categories (e.g., Dogs), 164 subjects (e.g., Labrador), and 183 properties (e.g., is a mammal). We then use TAXI to evaluate popular editors' categorical consistency, measuring how often editing a subject's category appropriately edits its properties. We find that 1) the editors achieve marginal, yet non-random consistency, 2) their consistency far underperforms human baselines, and 3) consistency is more achievable when editing atypical subjects Our code and data are available at https://github.com/derekpowell/taxi.