CoChat
Sign Up
Cited by 3Natural Language Processing TechniquesTopic ModelingInformation Retrieval and Search Behavior

Teach Better or Show Smarter? On Instructions and Exemplars in Automatic Prompt Optimization

Open access
Wan, Xingchen, Sun, Ruoxi, Nakhost, Hootan, Arik, Sercan·2024·Advances in Neural Information Processing Systems 37
DOI 10.52202/079017-1855
Open this paper in CoChat — free accountOpen PDF

Abstract

Large language models have demonstrated remarkable capabilities, but their performance is heavily reliant on effective prompt engineering. Automatic prompt optimization (APO) methods are designed to automate this and can be broadly categorized into those targeting instructions (instruction optimization, IO) vs. those targeting exemplars (exemplar optimization, EO). Despite their shared objective, these have evolved rather independently, with IO receiving more research attention recently. This paper seeks to bridge this gap by comprehensively comparing the performance of representative IO and EO techniques both isolation and combination on a diverse set of challenging tasks. Our findings reveal that intelligently reusing model-generated input-output pairs obtained from evaluating prompts on the validation set as exemplars, consistently improves performance on top of IO methods but is currently under-investigated. We also find that despite the recent focus on IO, how we select exemplars can outweigh how we optimize instructions, with EO strategies as simple as random search outperforming state-of-the-art IO methods with seed instructions without any optimization. Moreover, we observe a synergy between EO and IO, with optimal combinations surpassing the individual contributions. We conclude that studying exemplar optimization both as a standalone method and its optimal combination with instruction optimization remain a crucial aspect of APO and deserve greater consideration in future research, even in the era of highly capable instruction-following models.

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.52202/079017-1855

Cites

Statistics

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

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?

  • GEPA's prompts frequently contain detailed declarative instructions for completing the task, in contrast to prior findings where instruction optimization yielded improvements primarily through quasi-exemplars.

    “Notably, in contrast to prior findings where instruction optimization yielded improvements primarily through quasi-exemplars (Wan et al., 2024), GEPA's prompts frequently contain detailed declarative instructions for completing the task, as illustrated in Figure 2.”

    From GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learning (2025) · cites this paper as Wan et al., 2024

  • Our results suggest that instructions generated by reflective prompt evolution can achieve stronger generalization as well as improved overall performance.

    “While Wan et al. (2024) previously observed that exemplars tend to generalize better, our results suggest that instructions generated by reflective prompt evolution can achieve stronger generalization as well as improved overall performance.”

    From GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learning (2025) · cites this paper as Wan et al. (2024)

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Cites

No references could be resolved yet.

Research works printed but not linked: 67

  1. Achiam, J., Adler, S., Agarwal, S., Ahmad, L., Akkaya, I., Aleman, F. L., Almeida, D., Altenschmidt, J., Altman, S., Anadkat, S., et al. GPT-4 technical report. arXiv preprint arXiv:2303.08774, 2023. · arXiv
  2. Agarwal, E., Dani, V., Ganu, T., and Nambi, A. Promptwizard: Task-aware agent-driven prompt optimization framework. arXiv preprint arXiv:2405.18369, 2024a. · arXiv
  3. Agarwal, R., Singh, A., Zhang, L. M., Bohnet, B., Chan, S., Anand, A., Abbas, Z., Nova, A., Co-Reyes, J. D., Chu, E., et al. Many-shot in-context learning. arXiv preprint arXiv:2404.11018, 2024b. · arXiv
  4. Anil, R., Dai, A. M., Firat, O., Johnson, M., Lepikhin, D., Passos, A., Shakeri, S., Taropa, E., Bailey, P., Chen, Z., et al. Palm 2 technical report. arXiv preprint arXiv:2305.10403, 2023. · arXiv
  5. Baptista, R. and Poloczek, M. Bayesian optimization of combinatorial structures. In International Conference on Machine Learning, pp. 462-471. PMLR, 2018.
  6. Brown, T., Mann, B., Ryder, N., Subbiah, M., Kaplan, J. D., Dhariwal, P., Neelakantan, A., Shyam, P., Sastry, G., Askell, A., et al. Language models are few-shot learners. Advances in neural information processing systems, 33:1877-1901, 2020.
  7. Chen, L., Chen, J., Goldstein, T., Huang, H., and Zhou, T. Instructzero: Efficient instruction optimiza- tion for black-box large language models. arXiv preprint arXiv:2306.03082, 2023. · arXiv
  8. Chen, Y., Zhao, C., Yu, Z., McKeown, K., and He, H. On the relation between sensitivity and accuracy in in-context learning. arXiv preprint arXiv:2209.07661, 2022. · arXiv
  9. Chowdhery, A., Narang, S., Devlin, J., Bosma, M., Mishra, G., Roberts, A., Barham, P., Chung, H. W., Sutton, C., Gehrmann, S., et al. Palm: Scaling language modeling with pathways. Journal of Machine Learning Research, 24(240):1-113, 2023.
  10. Coulom, R. Efficient selectivity and backup operators in monte-carlo tree search. In International conference on computers and games, pp. 72-83. Springer, 2006.
  11. Daulton, S., Wan, X., Eriksson, D., Balandat, M., Osborne, M. A., and Bakshy, E. Bayesian optimization over discrete and mixed spaces via probabilistic reparameterization. Advances in Neural Information Processing Systems, 35:12760-12774, 2022.
  12. Deng, M., Wang, J., Hsieh, C.-P., Wang, Y., Guo, H., Shu, T., Song, M., Xing, E., and Hu, Z. RLPrompt: Optimizing discrete text prompts with reinforcement learning. In Goldberg, Y., Kozareva, Z., and Zhang, Y. (eds.), Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pp. 3369-3391, Abu Dhabi, United Arab Emirates, December 2022. Association for Computational Linguistics. doi: 10.18653/v1/2022.emnlp-main.222. URL https://aclanthology.org/2022.emnlp-main.222. · DOI
  13. Do, X. L., Zhao, Y., Brown, H., Xie, Y., Zhao, J. X., Chen, N. F., Kawaguchi, K., Xie, M. Q., and He, J. Prompt optimization via adversarial in-context learning. arXiv preprint arXiv:2312.02614, 2023. · arXiv
  14. Fernando, C., Banarse, D., Michalewski, H., Osindero, S., and Rocktäschel, T. Promptbreeder: Self-referential self-improvement via prompt evolution. arXiv preprint arXiv:2309.16797, 2023. · arXiv
  15. Gemini team, Anil, R., Borgeaud, S., Wu, Y., Alayrac, J.-B., Yu, J., Soricut, R., Schalkwyk, J., Dai, A. M., Hauth, A., et al. Gemini: a family of highly capable multimodal models. arXiv preprint arXiv:2312.11805, 2023. · arXiv
  16. Guo, Q., Wang, R., Guo, J., Li, B., Song, K., Tan, X., Liu, G., Bian, J., and Yang, Y. Connecting large language models with evolutionary algorithms yields powerful prompt optimizers. International Conference on Learning Representations, 2024.
  17. Hendrycks, D., Burns, C., Basart, S., Zou, A., Mazeika, M., Song, D., and Steinhardt, J. Measuring massive multitask language understanding. arXiv preprint arXiv:2009.03300, 2020. · arXiv
  18. Honovich, O., Shaham, U., Bowman, S. R., and Levy, O. Instruction induction: From few examples to natural language task descriptions. In Rogers, A., Boyd-Graber, J., and Okazaki, N. (eds.), Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pp. 1935-1952, Toronto, Canada, July 2023. Association for Computational Linguistics. doi: 10.18653/v1/2023.acl-long.108. URL https://aclanthology.org/2023.acl-long.108. · DOI
  19. Hsieh, C.-J., Si, S., Yu, F. X., and Dhillon, I. S. Automatic engineering of long prompts. arXiv preprint arXiv:2311.10117, 2023. · arXiv
  20. Khattab, O., Santhanam, K., Li, X. L., Hall, D., Liang, P., Potts, C., and Zaharia, M. Demonstrate-search- predict: Composing retrieval and language models for knowledge-intensive nlp. arXiv preprint arXiv:2212.14024, 2022. · arXiv
  21. Khattab, O., Singhvi, A., Maheshwari, P., Zhang, Z., Santhanam, K., Vardhamanan, S., Haq, S., Sharma, A., Joshi, T. T., Moazam, H., et al. DSPy: Compiling declarative language model calls into self-improving pipelines. International Conference on Learning Representations, 2024.
  22. Kim, H. J., Cho, H., Kim, J., Kim, T., Yoo, K. M., and Lee, S.-g. Self-generated in-context learn- ing: Leveraging auto-regressive language models as a demonstration generator. arXiv preprint arXiv:2206.08082, 2022. · arXiv
  23. Kojima, T., Gu, S. S., Reid, M., Matsuo, Y., and Iwasawa, Y. Large language models are zero-shot reasoners. Advances in neural information processing systems, 35:22199-22213, 2022.
  24. Lee, J., Dai, Z., Ren, X., Chen, B., Cer, D., Cole, J. R., Hui, K., Boratko, M., Kapadia, R., Ding, W., et al. Gecko: Versatile text embeddings distilled from large language models. arXiv preprint arXiv:2403.20327, 2024. · arXiv
  25. Lin, X., Wu, Z., Dai, Z., Hu, W., Shu, Y., Ng, S.-K., Jaillet, P., and Low, B. K. H. Use your instinct: Instruc- tion optimization using neural bandits coupled with transformers. arXiv preprint arXiv:2310.02905, 2023. · arXiv
  26. Liu, P., Yuan, W., Fu, J., Jiang, Z., Hayashi, H., and Neubig, G. Pre-train, prompt, and predict: A systematic survey of prompting methods in natural language processing. ACM Computing Surveys, 55(9):1-35, 2023a.
  27. Liu, S., Chen, C., Qu, X., Tang, K., and Ong, Y.-S. Large language models as evolutionary optimizers. arXiv preprint arXiv:2310.19046, 2023b. · arXiv
  28. Lu, Y., Bartolo, M., Moore, A., Riedel, S., and Stenetorp, P. Fantastically ordered prompts and where to find them: Overcoming few-shot prompt order sensitivity. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pp. 8086-8098, 2022.
  29. Margatina, K., Schick, T., Aletras, N., and Dwivedi-Yu, J. Active learning principles for in-context learning with large language models. In Bouamor, H., Pino, J., and Bali, K. (eds.), Findings of the Association for Computational Linguistics: EMNLP 2023, pp. 5011-5034, Singapore, December 2023. Association for Computational Linguistics. doi: 10.18653/v1/2023.findings-emnlp.334. URL https://aclanthology.org/2023.findings-emnlp.334. · DOI
  30. Mizrahi, M., Kaplan, G., Malkin, D., Dror, R., Shahaf, D., and Stanovsky, G. State of what art? a call for multi-prompt llm evaluation. arXiv preprint arXiv:2401.00595, 2024. · arXiv
  31. Nguyen, T. and Wong, E. In-context example selection with influences. arXiv preprint arXiv:2302.11042, 2023. · arXiv
  32. Opsahl-Ong, K., Ryan, M. J., Purtell, J., Broman, D., Potts, C., Zaharia, M., and Khattab, O. Opti- mizing instructions and demonstrations for multi-stage language model programs. arXiv preprint arXiv:2406.11695, 2024. · arXiv
  33. Perez, E., Kiela, D., and Cho, K. True few-shot learning with language models. Advances in neural information processing systems, 34:11054-11070, 2021.
  34. Prasad, A., Hase, P., Zhou, X., and Bansal, M. GrIPS: Gradient-free, edit-based instruction search for prompting large language models. In Vlachos, A. and Augenstein, I. (eds.), Proceedings of the 17th Conference of the European Chapter of the Association for Computational Linguistics, pp. 3845-3864, Dubrovnik, Croatia, May 2023. Association for Computational Linguistics. doi: 10.18653/v1/2023. eacl-main.277. URL https://aclanthology.org/2023.eacl-main.277. · DOI · Paper page
  35. Pryzant, R., Iter, D., Li, J., Lee, Y., Zhu, C., and Zeng, M. Automatic prompt optimization with "gradient descent" and beam search. In Bouamor, H., Pino, J., and Bali, K. (eds.), Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pp. 7957-7968, Singapore, December 2023. Association for Computational Linguistics. doi: 10.18653/v1/2023.emnlp-main. 494. URL https://aclanthology.org/2023.emnlp-main.494. · DOI · Paper page
  36. Reid, M., Savinov, N., Teplyashin, D., Lepikhin, D., Lillicrap, T., Alayrac, J.-b., Soricut, R., Lazaridou, A., Firat, O., Schrittwieser, J., et al. Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context. arXiv preprint arXiv:2403.05530, 2024. · arXiv
  37. Rubin, O., Herzig, J., and Berant, J. Learning to retrieve prompts for in-context learning. In Carpuat, M., de Marneffe, M.-C., and Meza Ruiz, I. V. (eds.), Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pp. 2655-2671, Seattle, United States, July 2022. Association for Computational Linguistics. doi: 10.18653/v1/2022.naacl-main.191. URL https://aclanthology.org/2022.naacl-main. 191. · DOI · Paper page
  38. Sahoo, P., Singh, A. K., Saha, S., Jain, V., Mondal, S., and Chadha, A. A systematic survey of prompt en- gineering in large language models: Techniques and applications. arXiv preprint arXiv:2402.07927, 2024. · arXiv
  39. Sclar, M., Choi, Y., Tsvetkov, Y., and Suhr, A. Quantifying language models' sensitivity to spurious features in prompt design or: How i learned to start worrying about prompt formatting. arXiv preprint arXiv:2310.11324, 2023. · arXiv
  40. Shin, T., Razeghi, Y., Logan IV, R. L., Wallace, E., and Singh, S. Autoprompt: Eliciting knowledge from language models with automatically generated prompts. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pp. 4222-4235, 2020.
  41. Stanić, A., Caelles, S., and Tschannen, M. Towards truly zero-shot compositional visual reasoning with llms as programmers. arXiv preprint arXiv:2401.01974, 2024. · arXiv
  42. Sun, H., Li, X., Xu, Y., Homma, Y., Cao, Q., Wu, M., Jiao, J., and Charles, D. Autohint: Automatic prompt optimization with hint generation. arXiv preprint arXiv:2307.07415, 2023. · arXiv
  43. Suzgun, M., Scales, N., Schärli, N., Gehrmann, S., Tay, Y., Chung, H. W., Chowdhery, A., Le, Q. V., Chi, E. H., Zhou, D., et al. Challenging big-bench tasks and whether chain-of-thought can solve them. arXiv preprint arXiv:2210.09261, 2022. · arXiv
  44. Wan, X., Nguyen, V., Ha, H., Ru, B., Lu, C., and Osborne, M. A. Think global and act local: Bayesian optimisation over high-dimensional categorical and mixed search spaces. In International Conference on Machine Learning, pp. 10663-10674. PMLR, 2021.
  45. Wan, X., Sun, R., Dai, H., Arik, S., and Pfister, T. Better zero-shot reasoning with self-adaptive prompting. In Rogers, A., Boyd-Graber, J., and Okazaki, N. (eds.), Findings of the Associa- tion for Computational Linguistics: ACL 2023, pp. 3493-3514, Toronto, Canada, July 2023a. Association for Computational Linguistics. doi: 10.18653/v1/2023.findings-acl.216. URL https://aclanthology.org/2023.findings-acl.216. · DOI
  46. Wan, X., Sun, R., Nakhost, H., Dai, H., Eisenschlos, J., Arik, S., and Pfister, T. Universal self-adaptive prompting. In Bouamor, H., Pino, J., and Bali, K. (eds.), Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pp. 7437-7462, Singapore, December 2023b. Association for Computational Linguistics. doi: 10.18653/v1/2023.emnlp-main.461. URL https: //aclanthology.org/2023.emnlp-main.461. · DOI
  47. Wang, R., An, S., Cheng, M., Zhou, T., Hwang, S. J., and Hsieh, C.-J. Mixture-of-experts in prompt optimization. International Conference on Machine Learning, 2024a.
  48. Wang, X., Li, C., Wang, Z., Bai, F., Luo, H., Zhang, J., Jojic, N., Xing, E. P., and Hu, Z. Promptagent: Strategic planning with language models enables expert-level prompt optimization. International Conference on Learning Representations, 2024b.
  49. Wei, J., Bosma, M., Zhao, V., Guu, K., Yu, A. W., Lester, B., Du, N., Dai, A. M., and Le, Q. V. Finetuned language models are zero-shot learners. In International Conference on Learning Representations, 2022a. URL https://openreview.net/forum?id=gEZrGCozdqR.
  50. Wei, J., Wang, X., Schuurmans, D., Bosma, M., Xia, F., Chi, E., Le, Q. V., Zhou, D., et al. Chain-of- thought prompting elicits reasoning in large language models. Advances in neural information processing systems, 35:24824-24837, 2022b.
  51. Wu, Z., Wang, Y., Ye, J., and Kong, L. Self-adaptive in-context learning: An information compression perspective for in-context example selection and ordering. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pp. 1423-1436, 2023.
  52. Wu, Z., Lin, X., Dai, Z., Hu, W., Shu, Y., Ng, S.-K., Jaillet, P., and Low, B. K. H. Prompt optimization with ease? efficient ordering-aware automated selection of exemplars. arXiv preprint arXiv:2405.16122, 2024. · arXiv
  53. Xu, H., Chen, Y., Du, Y., Shao, N., Wang, Y., Li, H., and Yang, Z. Gps: Genetic prompt search for efficient few-shot learning. arXiv preprint arXiv:2210.17041, 2022. · arXiv
  54. Yang, C., Wang, X., Lu, Y., Liu, H., Le, Q. V., Zhou, D., and Chen, X. Large language models as optimizers. In International Conference on Learning Representations, 2024.
  55. Ye, J., Wu, Z., Feng, J., Yu, T., and Kong, L. Compositional exemplars for in-context learning. In International Conference on Machine Learning, pp. 39818-39833.
  56. PMLR, 2023a.
  57. Ye, Q., Axmed, M., Pryzant, R., and Khani, F. Prompt engineering a prompt engineer. arXiv preprint arXiv:2311.05661, 2023b. · arXiv
  58. Zhang, T., Wang, X., Zhou, D., Schuurmans, D., and Gonzalez, J. E. Tempera: Test-time prompting via reinforcement learning. International Conference on Learning Representations, 2023a.
  59. Zhang, Y., Feng, S., and Tan, C. Active example selection for in-context learning. In Goldberg, Y., Kozareva, Z., and Zhang, Y. (eds.), Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pp. 9134-9148, Abu Dhabi, United Arab Emirates, December 2022. Association for Computational Linguistics. doi: 10.18653/v1/2022.emnlp-main.622. URL https://aclanthology.org/2022.emnlp-main.622. · DOI
  60. Zhang, Y., Zhou, K., and Liu, Z. What makes good examples for visual in-context learning? Advances in Neural Information Processing Systems, 36, 2024.
  61. Zhang, Z., Zhang, A., Li, M., and Smola, A. Automatic chain of thought prompting in large language models. International Conference on Learning Representations, 2023b.
  62. Zhao, Z., Wallace, E., Feng, S., Klein, D., and Singh, S. Calibrate before use: Improving few-shot performance of language models. In International conference on machine learning, pp. 12697-12706.
  63. PMLR, 2021.
  64. Zhou, H., Wan, X., Vulić, I., and Korhonen, A. Survival of the most influential prompts: Efficient black- box prompt search via clustering and pruning. In Bouamor, H., Pino, J., and Bali, K. (eds.), Findings of the Association for Computational Linguistics: EMNLP 2023, pp. 13064-13077, Singapore, December 2023a. Association for Computational Linguistics. doi: 10.18653/v1/2023.findings-emnlp.870. URL https://aclanthology.org/2023.findings-emnlp.870. · DOI
  65. Zhou, H., Wan, X., Liu, Y., Collier, N., Vulić, I., and Korhonen, A. Fairer preferences elicit improved human-aligned large language model judgments. arXiv preprint arXiv:2406.11370, 2024a. · arXiv
  66. Zhou, H., Wan, X., Proleev, L., Mincu, D., Chen, J., Heller, K., and Roy, S. Batch calibration: Rethinking calibration for in-context learning and prompt engineering. International Conference on Learning Representations (ICLR), 2024b.
  67. Zhou, H., Wan, X., Vulić, I., and Korhonen, A. Autopeft: Automatic configuration search for parameter- efficient fine-tuning. Transactions of the Association for Computational Linguistics, 12:525-542, 2024c.

Cited by

Sorted by citations

Authors
Wan, X., Sun, R., Nakhost, H., & Arik, S.
ORCID
Venue
Advances in Neural Information Processing Systems 37
Publisher
Neural Information Processing Systems Foundation, Inc. (NeurIPS)
Published
2024
Pages
58174-58244
ISSN
1049-5258 (ISSN-L)
DOI
10.52202/079017-1855
OpenAlex
W7103757416
Cited by
3
References
0
Citations by year
2026: 1 · 2025: 2
FWCI
0.87
Type
conference paper
Language
en
Open access
Locations
  • neural information processing systems (publishedVersion)
  • arXiv (Cornell University) (submittedVersion)
  • arXiv (Cornell University)
Topics
  • Physical Sciences › Computer Science › Artificial Intelligence › Natural Language Processing Techniques
  • Physical Sciences › Computer Science › Artificial Intelligence › Topic Modeling
  • Physical Sciences › Computer Science › Information Systems › Information Retrieval and Search Behavior
Keywords
prompt optimization, instruction, exemplars, large language models

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