Salvato in:
| Autori principali: | , , , , , , , , , , , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2405.09605 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866911037319544832 |
|---|---|
| author | Ivanova, Anna A. Sathe, Aalok Lipkin, Benjamin Kumar, Unnathi Radkani, Setayesh Clark, Thomas H. Kauf, Carina Hu, Jennifer Pramod, R. T. Grand, Gabriel Paulun, Vivian Ryskina, Maria Akyürek, Ekin Wilcox, Ethan Rashid, Nafisa Choshen, Leshem Levy, Roger Fedorenko, Evelina Tenenbaum, Joshua Andreas, Jacob |
| author_facet | Ivanova, Anna A. Sathe, Aalok Lipkin, Benjamin Kumar, Unnathi Radkani, Setayesh Clark, Thomas H. Kauf, Carina Hu, Jennifer Pramod, R. T. Grand, Gabriel Paulun, Vivian Ryskina, Maria Akyürek, Ekin Wilcox, Ethan Rashid, Nafisa Choshen, Leshem Levy, Roger Fedorenko, Evelina Tenenbaum, Joshua Andreas, Jacob |
| contents | The ability to build and reason about models of the world is essential for situated language understanding. But evaluating world modeling capabilities in modern AI systems -- especially those based on language models -- has proven challenging, in large part because of the difficulty of disentangling conceptual knowledge about the world from knowledge of surface co-occurrence statistics. This paper presents Elements of World Knowledge (EWoK), a framework for evaluating language models' understanding of the conceptual knowledge underlying world modeling. EWoK targets specific concepts from multiple knowledge domains known to be important for world modeling in humans, from social interactions (help, deceive) to spatial relations (left, right). Objects, agents, and locations in the items can be flexibly filled in, enabling easy generation of multiple controlled datasets. We then introduce EWoK-core-1.0, a dataset of 4,374 items covering 11 world knowledge domains. We evaluate 20 open-weights large language models (1.3B--70B parameters) and compare them with human performance. All tested models perform worse than humans, with results varying drastically across domains. Performance on social interactions and social properties was highest and performance on physical relations and spatial relations was lowest. Overall, this dataset highlights simple cases where even large models struggle and presents rich avenues for targeted research on LLM world modeling capabilities. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2405_09605 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Elements of World Knowledge (EWoK): A Cognition-Inspired Framework for Evaluating Basic World Knowledge in Language Models Ivanova, Anna A. Sathe, Aalok Lipkin, Benjamin Kumar, Unnathi Radkani, Setayesh Clark, Thomas H. Kauf, Carina Hu, Jennifer Pramod, R. T. Grand, Gabriel Paulun, Vivian Ryskina, Maria Akyürek, Ekin Wilcox, Ethan Rashid, Nafisa Choshen, Leshem Levy, Roger Fedorenko, Evelina Tenenbaum, Joshua Andreas, Jacob Computation and Language Artificial Intelligence Machine Learning The ability to build and reason about models of the world is essential for situated language understanding. But evaluating world modeling capabilities in modern AI systems -- especially those based on language models -- has proven challenging, in large part because of the difficulty of disentangling conceptual knowledge about the world from knowledge of surface co-occurrence statistics. This paper presents Elements of World Knowledge (EWoK), a framework for evaluating language models' understanding of the conceptual knowledge underlying world modeling. EWoK targets specific concepts from multiple knowledge domains known to be important for world modeling in humans, from social interactions (help, deceive) to spatial relations (left, right). Objects, agents, and locations in the items can be flexibly filled in, enabling easy generation of multiple controlled datasets. We then introduce EWoK-core-1.0, a dataset of 4,374 items covering 11 world knowledge domains. We evaluate 20 open-weights large language models (1.3B--70B parameters) and compare them with human performance. All tested models perform worse than humans, with results varying drastically across domains. Performance on social interactions and social properties was highest and performance on physical relations and spatial relations was lowest. Overall, this dataset highlights simple cases where even large models struggle and presents rich avenues for targeted research on LLM world modeling capabilities. |
| title | Elements of World Knowledge (EWoK): A Cognition-Inspired Framework for Evaluating Basic World Knowledge in Language Models |
| topic | Computation and Language Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2405.09605 |