The DOI matches its Crossref record; 16 of the 16 references it cites were found in a registry or an index.
In-Context Learning Creates Task Vectors
Abstract
In-context learning (ICL) in Large Language Models (LLMs) has emerged as a powerful new learning paradigm.However, its underlying mechanism is still not well understood.In particular, it is challenging to map it to the "standard" machine learning framework, where one uses a training set S to find a best-fitting function f (x) in some hypothesis class.Here we make progress on this problem by showing that the functions learned by ICL often have a very simple structure: they correspond to the transformer LLM whose only inputs are the query x and a single "task vector" calculated from the training set.Thus, ICL can be seen as compressing S into a single task vector θ(S) and then using this task vector to modulate the transformer to produce the output.We support the above claim via comprehensive experiments across a range of models and tasks. 1
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
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?
In the citing paper
“Prior work has demonstrated that different prompting strategies can lead to measurably different activation patterns (Wei et al. 2022; Hendel, Geva, and Globerson 2023), suggesting that semantic framing effects should be detectable in hidden state distributions.”
From When Truth Is Overridden: Uncovering the Internal Origins of Sycophancy in Large Language Models (2026) · cites this paper as Hendel, Geva, and Globerson 2023
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Language Models are Few-Shot Learners
Brown, T. B., Mann, B., Ryder, N., Subbiah, M., Kaplan, J., Dhariwal, P., Neelakantan, A., Shyam, P., Sastry, G., Askell, A., et al. (2020) · arXiv · doi:10.48550/arxiv.2005.14165 · Cited by 23,372
LLaMA: Open and Efficient Foundation Language Models
Touvron, H., Lavril, T., Izacard, G., Martinet, X., Lachaux, M.-A., Lacroix, T., Rozière, B., Goyal, N., Hambro, E., Azhar, F., et al. (2023) · arXiv · doi:10.48550/arxiv.2302.13971 · Cited by 4,602
Liu, P., Yuan, W., Fu, J., Jiang, Z., Hayashi, H., & Neubig, G. (2023) · ACM Computing Surveys · doi:10.1145/3560815 · Cited by 4,361
Liu, P., Yuan, W., Fu, J., Jiang, Z., Hayashi, H., & Neubig, G. (2021) · arXiv · doi:10.48550/arxiv.2107.13586 · Cited by 4,352
Understanding Machine Learning: From Theory to Algorithms
Shalev-Shwartz, S., & Ben-David, S. (2014) · doi:10.1017/cbo9781107298019 · Cited by 3,115
The Power of Scale for Parameter-Efficient Prompt Tuning
Lester, B., Al-Rfou, R., & Constant, N. (2021) · Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/2021.emnlp-main.243 · Cited by 3,096
The Power of Scale for Parameter-Efficient Prompt Tuning
Lester, B., Al-Rfou, R., & Constant, N. (2021) · arXiv · doi:10.48550/arxiv.2104.08691 · Cited by 3,092
In-context Learning and Induction Heads
Olsson, C., Elhage, N., Nanda, N., Joseph, N., DasSarma, N., Henighan, T., Mann, B., Askell, A., Bai, Y., Chen, A., et al. (2022) · arXiv · doi:10.48550/arxiv.2209.11895 · Cited by 1,170
A Survey on In-context Learning
Dong, Q., Li, L., Dai, D., Zheng, C., Ma, J., Li, R., Xia, H., Xu, J., Wu, Z., Liu, T., et al. (2023) · arXiv · doi:10.48550/arxiv.2301.00234 · Cited by 248
Locating and Editing Factual Associations in GPT
Meng, K., Bau, D., Andonian, A., & Belinkov, Y. (2022) · arXiv · doi:10.48550/arxiv.2202.05262 · Cited by 230
Cited by
The mystery of In-context learning: a comprehensive survey on interpretation and analysis
Zhou, Y., Li, J., Xiang, Y., Yan, H., Gui, L., & He, Y. (2023) · arXiv · doi:10.48550/arxiv.2311.00237 · Cited by 33
The Geometry of Concepts: Sparse Autoencoder Feature Structure
Li, Y., Michaud, E. J., Baek, D. D., Engels, J., Sun, X., & Tegmark, M. (2024) · arXiv · doi:10.48550/arxiv.2410.19750 · Cited by 14
Yager, K. G. (2024) · Digital Discovery · doi:10.1039/d4dd00178h · Cited by 12
Improving Instruction-Following in Language Models through Activation Steering
Stolfo, A., Balachandran, V., Yousefi, S., Horvitz, E., & Nushi, B. (2024) · arXiv · doi:10.48550/arxiv.2410.12877 · Cited by 10
When Truth Is Overridden: Uncovering the Internal Origins of Sycophancy in Large Language Models
Wang, K., Li, J., Yang, S., Zhang, Z., & Wang, D. (2026) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v40i39.40645 · Cited by 8
Metadata
- Authors
- Hendel, R., Geva, M., & Globerson, A.
- Venue
- Findings of the Association for Computational Linguistics: EMNLP 2023
- Published
- 2023
- Pages
- 9318-9333
- Cited by
- 44
- References
- 16
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
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: In-Context Learning Creates Task Vectors. https://app.cochat.ai/papers/10.18653/v1/2023.findings-emnlp.624
- The paper
- Hendel, R., Geva, M., & Globerson, A. (2023). In-Context Learning Creates Task Vectors. Findings of the Association for Computational Linguistics: EMNLP 2023. https://doi.org/10.18653/v1/2023.findings-emnlp.624
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.