Elamur Research
☰
home / papers / 2023 / Sparks of Artificial General Intelligence: Early experiments with GPT-4

Sparks of Artificial General Intelligence: Early experiments with GPT-4

Written for the Research library (claude-opus-5-5) from the paper text. Original — arxiv.org/pdf/2303.12712

Summary

The paper asks whether a large language model can show broad, general-purpose intelligence rather than skill at one narrow task. A Microsoft Research team had early access to a pre-release, text-only build of OpenAI's GPT-4. They probed it qualitatively with plain natural-language prompts and no special prompting techniques. The tests cover interdisciplinary composition (for example, a poem proving that infinitely many primes exist, or a unicorn drawn in TikZ), coding, mathematics, tool use, embodied interaction, theory of mind, explainability, and discriminative tasks such as PII detection and fact-checking. ChatGPT serves as the main baseline. The authors report that GPT-4 often comes close to human-level performance and clearly outperforms ChatGPT. On that basis they argue it can be read as an early, incomplete form of AGI. They also document weaknesses: hallucinations, weak planning caused by autoregressive next-word generation, sensitivity to prompt wording, and biases inherited from training data. The work matters as an influential, though purely observational, argument for treating LLMs as general systems and for exploring designs beyond next-word prediction.

Key points

  • The subject is an early, non-multimodal development version of GPT-4, which the authors assessed only through natural-language interaction. ChatGPT, a fine-tuned improved GPT-3, is the reference model.
  • Demonstrations span vision (via code-generated images), music, coding (including LeetCode-style challenges and comparison with human performance), mathematics, medicine, law, psychology, tool use, and text-based games.
  • A dedicated section on limitations of the autoregressive architecture shows failures in planning, both in arithmetic and reasoning and in text generation.
  • The authors suggest several extensions to next-word prediction:
    • calls to external tools such as a calculator, search, or code execution
    • a "slow-thinking" supervisory mechanism
    • built-in long-term memory
    • hierarchical, top-down generation
  • The authors acknowledge the study is phenomenological. It does not explain why or how GPT-4 works, and it cannot separate flaws introduced by RLHF from flaws inherent to the architecture.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

Fields: Other

Cited by

865 papers of our library cite it (of the 140673 whose bibliographies we read); 65 of them have a page here, 800 not yet.

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.48550/arxiv.2303.12712 · arXiv 2303.12712

@misc{bubeck2023sparks,
  title = {{Sparks of Artificial General Intelligence: Early experiments with GPT-4}},
  author = {Bubeck, Sébastien and Chandrasekaran, Varun and Eldan, Ronen and Gehrke, Johannes and Horvitz, Eric and Kamar, Ece and Lee, Peter and Lee, Yin Tat and Li, Yuanzhi and Lundberg, Scott and Nori, Harsha and Palangi, Hamid and Ribeiro, Marco Tulio and Zhang, Yi},
  year = {2023},
  doi = {10.48550/arxiv.2303.12712},
  eprint = {2303.12712},
  archivePrefix = {arXiv},
  url = {https://doi.org/10.48550/arxiv.2303.12712}
}

References

86 references, 19 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Kwangjun Ahn, Sébastien Bubeck, Sinho Chewi, Yin Tat Lee, Felipe Suarez, and Yi Zhang. Learning threshold neurons via the “edge of stability”. arXiv preprint arXiv:2212.07469, 2022. — [arXiv:2212.07469](https://arxiv.org/abs/2212.07469) find
  2. Saleema Amershi, Dan Weld, Mihaela Vorvoreanu, Adam Fourney, Besmira Nushi, Penny Col-lisson, Jina Suh, Shamsi Iqbal, Paul N Bennett, Kori Inkpen, Jaime Teevan, Ruth Kikin-Gil, and Eric Horvitz. Guidelines for human-AI interaction. In Proceedings of the 2019 CHI Conference on Human Factors in Comput… find
  3. Shikha Bordia and Samuel R Bowman. Identifying and reducing gender bias in word-level language models. arXiv preprint arXiv:1904.03035, 2019. find our card
  4. Su Lin Blodgett, Solon Barocas, Hal Daumé III, and Hanna Wallach. Language (technology) is power: A critical survey of” bias” in nlp. arXiv preprint arXiv:2005.14050, 2020. find our card
  5. Simon Baron-Cohen, Alan M Leslie, and Uta Frith. Does the autistic child have a “theory of mind”? Cognition, 21(1):37–46, 1985. find
  6. Tolga Bolukbasi, Kai-Wei Chang, James Y Zou, Venkatesh Saligrama, and Adam T Kalai. Man is to computer programmer as woman is to homemaker? Debiasing word embeddings. Advances in neural information processing systems, 29, 2016. find our card
  7. Boaz Barak, Benjamin L. Edelman, Surbhi Goel, Sham M. Kakade, eran malach, and Cyril Zhang. Hidden progress in deep learning: SGD learns parities near the computational limit. In Advances in Neural Information Processing Systems, 2022. find
  8. Emily M Bender, Timnit Gebru, Angelina McMillan-Major, and Shmargaret Shmitchell. On the dangers of stochastic parrots: Can language models be too big? In Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency, pages 610–623, 2021. find
  9. Dan Bohus and Eric Horvitz. Models for multiparty engagement in open-world dialog. In Proceedings of the SIGDIAL 2009 Conference, The 10th Annual Meeting of the Special Interest Group on Discourse and Dialogue, page 10, 2009. find
  10. Michael Bommarito II and Daniel Martin Katz. Gpt takes the bar exam. arXiv preprint arXiv:2212.14402, 2022. — [arXiv:2212.14402](https://arxiv.org/abs/2212.14402) find
  11. Erik Brynjolfsson and Tom Mitchell. What can machine learning do? workforce implications. Science, 358(6370):1530–1534, 2017. find
  12. Tom Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared D Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, Sandhini Agarwal, Ariel Herbert-Voss, Gretchen Krueger, Tom Henighan, Rewon Child, Aditya Ramesh, Daniel Ziegler, Jeffrey Wu, Clemens Winter, Chri… find
  13. Gagan Bansal, Besmira Nushi, Ece Kamar, Daniel S Weld, Walter S Lasecki, and Eric Horvitz. Updates in human-ai teams: Understanding and addressing the performance/compatibility tradeoff. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 33, pages 2429–2437, 2019. find
  14. Gagan Bansal, Besmira Nushi, Ece Kamar, Eric Horvitz, and Daniel S Weld. Is the most accurate ai the best teammate? Optimizing AI for teamwork. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 35, pages 11405–11414, 2021. find
  15. Sebastien Bubeck and Mark Sellke. A universal law of robustness via isoperimetry. In M. Ranzato, A. Beygelzimer, Y. Dauphin, P.S. Liang, and J. Wortman Vaughan, editors, Advances in Neural Information Processing Systems, volume 34, pages 28811–28822. Curran Associates, Inc., 2021. find
  16. François Chollet. On the measure of intelligence. arXiv preprint arXiv:1911.01547, 2019. find
  17. Karl Cobbe, Vineet Kosaraju, Mohammad Bavarian, Mark Chen, Heewoo Jun, Lukasz Kaiser, Matthias Plappert, Jerry Tworek, Jacob Hilton, Reiichiro Nakano, et al. Training verifiers to solve math word problems. arXiv preprint arXiv:2110.14168, 2021. find our card
  18. Marc-Alexandre Côté, Akos Kádár, Xingdi Yuan, Ben Kybartas, Tavian Barnes, Emery Fine, James Moore, Matthew Hausknecht, Layla El Asri, Mahmoud Adada, et al. Textworld: A learning environment for text-based games. In Workshop on Computer Games, pages 41–75. Springer, 2018. find our card
  19. Mark Chen, Jerry Tworek, Heewoo Jun, Qiming Yuan, Henrique Ponde de Oliveira Pinto, Jared Kaplan, Harri Edwards, Yuri Burda, Nicholas Joseph, Greg Brockman, Alex Ray, Raul Puri, Gretchen Krueger, Michael Petrov, Heidy Khlaaf, Girish Sastry, Pamela Mishkin, Brooke Chan, Scott Gray, Nick Ryder, Mikhai… find our card
  20. Katherine M Collins, Catherine Wong, Jiahai Feng, Megan Wei, and Josh Tenenbaum. Structured, flexible, and robust: benchmarking and improving large language models towards more human-like behavior in out-of-distribution reasoning tasks. In Proceedings of the Annual Meeting of the Cognitive Science S… find
  21. Maria De-Arteaga, Alexey Romanov, Hanna Wallach, Jennifer Chayes, Christian Borgs, Alexandra Chouldechova, Sahin Geyik, Krishnaram Kenthapadi, and Adam Tauman Kalai. Bias in bios: A case study of semantic representation bias in a high-stakes setting. In proceedings of the Conference on Fairness, Acc… find
  22. Ernest Davis and Gary Marcus. Commonsense reasoning and commonsense knowledge in artificial intelligence. Communications of the ACM, 58(9):92–103, 2015. find
  23. Ronen Eldan and Ohad Shamir. The power of depth for feedforward neural networks. In 29th Annual Conference on Learning Theory, volume 49 of Proceedings of Machine Learning Research, pages 907–940. PMLR, 2016. find
  24. Samuel J Gershman, Eric J Horvitz, and Joshua B Tenenbaum. Computational rationality: A converging paradigm for intelligence in brains, minds, and machines. Science, 349(6245):273– 278, 2015. find
  25. David Gillborn. Softly, softly: Genetics, intelligence and the hidden racism of the new geneism. Journal of Education Policy, 31(4):365–388, 2016. find
  26. Ben Goertzel. Artificial general intelligence: concept, state of the art, and future prospects. Journal of Artificial General Intelligence, 5(1):1, 2014. find
  27. Tejas Gokhale, Hamid Palangi, Besmira Nushi, Vibhav Vineet, Eric Horvitz, Ece Kamar, Chitta Baral, and Yezhou Yang. Benchmarking spatial relationships in text-to-image generation. arXiv preprint arXiv:2212.10015, 2022. — [arXiv:2212.10015](https://arxiv.org/abs/2212.10015) find
  28. Connie Guglielmo. CNET is experimenting with an AI assist. Here’s why, January 2023. [Online; posted 16-January-2023]. find
  29. Eric Horvitz and Matthew Barry. Display of information for time-critical decision making. In Proceedings of the UAI, 1995. find
  30. Dan Hendrycks, Collin Burns, Saurav Kadavath, Akul Arora, Steven Basart, Eric Tang, Dawn Song, and Jacob Steinhardt. Measuring mathematical problem solving with the math dataset. NeurIPS, 2021. find our card
  31. Eric Horvitz. Principles of mixed-initiative user interfaces. In Proceedings of the SIGCHI conference on Human Factors in Computing Systems, pages 159–166, 1999. find
  32. Eric Horvitz. Reflections on challenges and promises of mixed-initiative interaction. AI Magazine, 28(2), 2007. find
  33. Eric Horvitz. On the horizon: Interactive and compositional deepfakes. In Proceedings of the 2022 International Conference on Multimodal Interaction, page 653–661. Association for Computing Machinery, 2022. find
  34. Eric Horvitz and Tim Paek. Complementary computing: Policies for transferring callers from dialog systems to human receptionists. User Modeling and User-Adapted Interaction, 17(1):159– 182, 2007. find
  35. Dirk Hovy and Shannon L Spruit. The social impact of natural language processing. In Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pages 591–598, 2016. find our card
  36. Samy Jelassi, Michael E Sander, and Yuanzhi Li. Vision transformers provably learn spatial structure. arXiv preprint arXiv:2210.09221, 2022. — [arXiv:2210.09221](https://arxiv.org/abs/2210.09221) find
  37. Daniel Kahneman. Thinking, fast and slow. macmillan, 2011. find
  38. Ece Kamar, Severin Hacker, and Eric Horvitz. Combining human and machine intelligence in large-scale crowdsourcing. In AAMAS, volume 12, pages 467–474, 2012. find
  39. Aitor Lewkowycz, Anders Andreassen, David Dohan, Ethan Dyer, Henryk Michalewski, Vinay Ramasesh, Ambrose Slone, Cem Anil, Imanol Schlag, Theo Gutman-Solo, et al. Solving quantitative reasoning problems with language models. arXiv preprint arXiv:2206.14858, 2022. find our card
  40. Bingbin Liu, Jordan T Ash, Surbhi Goel, Akshay Krishnamurthy, and Cyril Zhang. Transformers learn shortcuts to automata. arXiv preprint arXiv:2210.10749, 2022. find
  41. Robert K Lindsay, Bruce G Buchanan, Edward A Feigenbaum, and Joshua Lederberg. Dendral: A case study of the first expert system for scientific hypothesis formation. Artificial Intelligence, 61(2):209–261, 1993. find
  42. Yann LeCun. A path towards autonomous machine intelligence. Open Review, 2022. find
  43. Lauren Leffer. CNET is reviewing the accuracy of all its AI-written articles after multiple major corrections, January 2023. [Online; posted 17-January-2023]. find
  44. Shane Legg. Machine super intelligence. PhD thesis, Università della Svizzera italiana, 2008. find
  45. Douglas B. Lenat. Cyc: A large-scale investment in knowledge infrastructure. Communications fo the ACM, 38(11):33–38, nov 1995. find
  46. Shane Legg and Marcus Hutter. Universal intelligence: A definition of machine intelligence. Minds and machines, 17(4):391–444, 2007. find
  47. Stephanie Lin, Jacob Hilton, and Owain Evans. Truthfulqa: Measuring how models mimic human falsehoods. arXiv preprint arXiv:2109.07958, 2021. find our card
  48. Chin-Yew Lin. Rouge: A package for automatic evaluation of summaries. In Text summarization branches out, pages 74–81, 2004. find
  49. Himabindu Lakkaraju, Ece Kamar, Rich Caruana, and Eric Horvitz. Identifying unknown unknowns in the open world: Representations and policies for guided exploration. In Thirty-first AAAI conference on artificial intelligence, 2017. find
  50. Patrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, et al. Retrieval-augmented generation for knowledge-intensive nlp tasks. Advances in Neural Information Processing Systems, 33:9459–9474, 2020. find
  51. Hussein Mozannar, Gagan Bansal, Adam Fourney, and Eric Horvitz. Reading between the lines: Modeling user behavior and costs in AI-assisted programming. arXiv preprint arXiv:2210.14306, 2022. — [arXiv:2210.14306](https://arxiv.org/abs/2210.14306) find
  52. Kyle Mahowald, Anna A Ivanova, Idan A Blank, Nancy Kanwisher, Joshua B Tenenbaum, and Evelina Fedorenko. Dissociating language and thought in large language models: a cognitive perspective. arXiv preprint arXiv:2301.06627, 2023. find
  53. Shikhar Murty, Christopher D Manning, Scott Lundberg, and Marco Tulio Ribeiro. Fixing model bugs with natural language patches. arXiv preprint arXiv:2211.03318, 2022. find
  54. John McCarthy, Marvin L Minsky, Nathaniel Rochester, and Claude E Shannon. A proposal for the Dartmouth summer research project on artificial intelligence, August 31, 1955. AI magazine, 27(4):12–12, 2006. find
  55. Joshua Maynez, Shashi Narayan, Bernd Bohnet, and Ryan McDonald. On faithfulness and factuality in abstractive summarization. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 1906–1919, 2020. find
  56. Mehryar Mohri, Afshin Rostamizadeh, and Ameet Talwalkar. Foundations of Machine Learning. MIT press, 2018. find
  57. Reiichiro Nakano, Jacob Hilton, Suchir Balaji, Jeff Wu, Long Ouyang, Christina Kim, Christopher Hesse, Shantanu Jain, Vineet Kosaraju, William Saunders, et al. Webgpt: Browser-assisted question-answering with human feedback. arXiv preprint arXiv:2112.09332, 2021. find our card
  58. Helen Nissenbaum. Privacy in context. In Privacy in Context. Stanford University Press, 2009. + find
  59. Harsha Nori, Nicholas King, Scott Mayer McKinney, Dean Carignan, and Eric Horvitz. Capabilities of GPT-4 on medical challenge problems. arXiv preprint arXiv:2303.13375, 2023. find our card
  60. Erik Nijkamp, Bo Pang, Hiroaki Hayashi, Lifu Tu, Huan Wang, Yingbo Zhou, Silvio Savarese, and Caiming Xiong. Codegen: An open large language model for code with multi-turn program synthesis. arXiv preprint, 2022. find
  61. Allen Newell, John C Shaw, and Herbert A Simon. Report on a general problem solving program. In IFIP congress, volume 256, page 64. Pittsburgh, PA, 1959. find
  62. Chris Olah, Nick Cammarata, Ludwig Schubert, Gabriel Goh, Michael Petrov, and Shan Carter. Zoom in: An introduction to circuits. Distill, 5(3):e00024–001, 2020. find
  63. Catherine Olsson, Nelson Elhage, Neel Nanda, Nicholas Joseph, Nova DasSarma, Tom Henighan, Ben Mann, Amanda Askell, Yuntao Bai, Anna Chen, et al. In-context learning and induction heads. arXiv preprint arXiv:2209.11895, 2022. find
  64. The University of Michigan. Tanner Lecture on AI and Human Values by Eric Horvitz. https: //www.youtube.com/watch?v=vsewugyXYXI, November 2022. find
  65. OpenAI. Gpt-4 technical report, 2023. arXiv preprint arXiv:2303.08774 [cs.CL]. find our card
  66. Brad Payne. Privacy protection with ai: Survey of data-anonymization techniques. 2020. + find
  67. Ildikó Pilán, Pierre Lison, Lilja Øvrelid, Anthi Papadopoulou, David Sánchez, and Montserrat Batet. The text anonymization benchmark (tab): A dedicated corpus and evaluation framework for text anonymization. arXiv preprint arXiv:2202.00443, 2022. find
  68. Kishore Papineni, Salim Roukos, Todd Ward, and Wei-Jing Zhu. Bleu: a method for automatic evaluation of machine translation. In Proceedings of the 40th annual meeting of the Association for Computational Linguistics, pages 311–318, 2002. find our card
  69. Krishna Pillutla, Swabha Swayamdipta, Rowan Zellers, John Thickstun, Sean Welleck, Yejin Choi, and Zaid Harchaoui. Mauve: Measuring the gap between neural text and human text using divergence frontiers. In Advances in Neural Information Processing Systems, volume 34, pages 4816–4828, 2021. find our card
  70. Ramya Ramakrishnan, Ece Kamar, Besmira Nushi, Debadeepta Dey, Julie Shah, and Eric Horvitz. Overcoming blind spots in the real world: Leveraging complementary abilities for joint execution. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 33, pages 6137–6145, 2019. find
  71. Kristen Reeder and Hwan Lee. Impact of artificial intelligence on us medical students’ choice of radiology. Clinical Imaging, 81:67–71, 2022. find
  72. Howard J Ross. Everyday bias: Identifying and navigating unconscious judgments in our daily lives. Rowman & Littlefield, 2020. find
  73. Karan Singhal, Shekoofeh Azizi, Tao Tu, S Sara Mahdavi, Jason Wei, Hyung Won Chung, Nathan Scales, Ajay Tanwani, Heather Cole-Lewis, Stephen Pfohl, et al. Large language models encode clinical knowledge. arXiv preprint arXiv:2212.13138, 2022. find our card
  74. Bart Selman, Rodney A Brooks, Thomas Dean, Eric Horvitz, Tom M Mitchell, and Nils J Nilsson. Challenge problems for artificial intelligence. In Proceedings of the National Conference on Artificial Intelligence, pages 1340–1345, 1996. find
  75. Thibault Sellam, Dipanjan Das, and Ankur P Parikh. Bleurt: Learning robust metrics for text generation. arXiv preprint arXiv:2004.04696, 2020. find
  76. Dafna Shahaf and Eric Horvitz. Generalized task markets for human and machine computation. In Twenty-Fourth AAAI Conference on Artificial Intelligence, 2010. find
  77. Adish Singla, Eric Horvitz, Pushmeet Kohli, and Andreas Krause. Learning to hire teams. In Third AAAI Conference on Human Computation and Crowdsourcing, 2015. find
  78. Aarohi Srivastava, Abhinav Rastogi, Abhishek Rao, Abu Awal Md Shoeb, Abubakar Abid, Adam Fisch, Adam R Brown, Adam Santoro, Aditya Gupta, Adrià Garriga-Alonso, et al. Beyond the imitation game: Quantifying and extrapolating the capabilities of language models. arXiv preprint arXiv:2206.04615, 2022. find our card
  79. Shai Shalev-Shwartz and Shai Ben-David. Understanding machine learning: From theory to algorithms. Cambridge university press, 2014. find
  80. Luca Venturi, Afonso S Bandeira, and Joan Bruna. Spurious valleys in one-hidden-layer neural network optimization landscapes. Journal of Machine Learning Research, 20:133, 2019. find
  81. Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, L ukasz Kaiser, and Illia Polosukhin. Attention is all you need. In Advances in Neural Information Processing Systems, volume 30, 2017. find
  82. Henry M Wellman. The child’s theory of mind. The MIT Press, 1992. find
  83. Bryan Wilder, Eric Horvitz, and Ece Kamar. Learning to complement humans. In Proceedings of the AAAI Conference on Artificial Intelligence, 2020. find
  84. Jason Wei, Yi Tay, Rishi Bommasani, Colin Raffel, Barret Zoph, Sebastian Borgeaud, Dani Yogatama, Maarten Bosma, Denny Zhou, Donald Metzler, Ed H. Chi, Tatsunori Hashimoto, Oriol Vinyals, Percy Liang, Jeff Dean, and William Fedus. Emergent abilities of large language models. Transactions on Machine… find our card
  85. Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Ed Chi, Quoc Le, and Denny Zhou. Chain of thought prompting elicits reasoning in large language models. arXiv preprint arXiv:2201.11903, 2022. find our card
  86. Yi Zhang, Arturs Backurs, Sébastien Bubeck, Ronen Eldan, Suriya Gunasekar, and Tal Wagner. Unveiling transformers with lego: a synthetic reasoning task. arXiv preprint arXiv:2206.04301, 2022. find