mlsanityData to 14 Oct 2026

Boundless Socratic Learning with Language Games

arXiv 2411.16905 ↗First posted 25 Nov 2024cs.AI cs.CLcited by 12

Not on the map. Why?

Abstract (arXiv)

An agent trained within a closed system can master any desired capability, as long as the following three conditions hold: (a) it receives sufficiently informative and aligned feedback, (b) its coverage of experience/data is broad enough, and (c) it has sufficient capacity and resource. In this position paper, we justify these conditions, and consider what limitations arise from (a) and (b) in closed systems, when assuming that (c) is not a bottleneck. Considering the special case of agents with matching input and output spaces (namely, language), we argue that such pure recursive self-improvement, dubbed "Socratic learning", can boost performance vastly beyond what is present in its initial data or knowledge, and is only limited by time, as well as gradual misalignment concerns. Furthermore, we propose a constructive framework to implement it, based on the notion of language games.

Key facts. Cited by 12 papers in the map (data through 14 Oct 2026). Builds on 26 of its 58 references that are in the map. First posted on arXiv 25 Nov 2024.

Builds on

26 of 58 references.

References · 26

  1. 2023GPT-4 Technical Reportcited by 18,396
  2. 2022Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedbackcited by 3,720
  3. 2022Constitutional AI: Harmlessness from AI Feedbackcited by 3,050
  4. 2023Voyager: An Open-Ended Embodied Agent with Large Language Modelscited by 2,191
  5. 2023Improving Factuality and Reasoning in Language Models through Multiagent Debatecited by 1,784
  6. 2024The AI Scientist: Towards Fully Automated Open-Ended Scientific Discoverycited by 1,071
  7. 2023Encouraging Divergent Thinking in Large Language Models through Multi-Agent Debatecited by 1,011
  8. 2023Promptbreeder: Self-Referential Self-Improvement Via Prompt Evolutioncited by 551
  9. 2018AI safety via debatecited by 400
  10. 2024Continual Learning of Large Language Models: A Comprehensive Surveycited by 234
  11. 2017Deal or No Deal? End-to-End Learning for Negotiation Dialoguescited by 209
  12. 2023FANToM: A Benchmark for Stress-testing Machine Theory of Mind in Interactionscited by 171
14 more
  1. 2018Supervising strong learners by amplifying weak expertscited by 145
  2. 2024AutoDefense: Multi-Agent LLM Defense against Jailbreak Attackscited by 145
  3. 2023Generative agent-based modeling with actions grounded in physical, social, or digital space using Concordiacited by 126
  4. 2023Levels of AGI for Operationalizing Progress on the Path to AGIcited by 96
  5. 2024Open-Endedness is Essential for Artificial Superhuman Intelligencecited by 80
  6. 2019Autocurricula and the Emergence of Innovation from Social Interaction: A Manifesto for Multi-Agent Intelligence Researchcited by 61
  7. 2023Empirical Design in Reinforcement Learningcited by 55
  8. 2024Thousands of AI Authors on the Future of AIcited by 48
  9. 2024Learning Formal Mathematics From Intrinsic Motivationcited by 45
  10. 2024GAVEL: Generating Games Via Evolution and Language Modelscited by 24
  11. 2021When should agents explore?cited by 20
  12. 2022Eliminating Meta Optimization Through Self-Referential Meta Learningcited by 14
  13. 2024Three Dogmas of Reinforcement Learningcited by 14
  14. 2024Does Thought Require Sensory Grounding? From Pure Thinkers to Large Language Modelscited by 7

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 5; 2026: 7.

12 papers.

  1. 2025Absolute Zero: Reinforced Self-play Reasoning with Zero Datacited by 344
  2. 2025A Survey on Test-Time Scaling in Large Language Models: What, How, Where, and How Well?cited by 223
  3. 2025SPIRAL: Self-Play on Zero-Sum Games Incentivizes Reasoning via Multi-Agent Multi-Turn Reinforcement Learningcited by 100
  4. 2026Recursive Self-Improvement in AI: From Bounded Self-Refinement to Autonomous Research Loopscited by 35
  5. 2025Multi-agent KTO: Reinforcing Strategic Interactions of Large Language Model in Language Gamecited by 10
  6. 2025Automated Capability Discovery via Foundation Model Self-Explorationcited by 10
  7. 2026Improving Interactive In-Context Learning from Natural Language Feedbackcited by 5
  8. 2026Position: Introspective Experience from Conversational Environments as a Path to Better Learningcited by 0
  9. 2026Playing with Words, Improving with Rewards: Training Language Models for Creative Associationcited by 0
  10. 2026An Information-Theoretic Definition for Open-Ended Learningcited by 0
  11. 2026Hallucination Self-Play: Bootstrapping Reinforced Detector via Evolved Generatorcited by 0
  12. 2026Structured Synthetic Reasoning Data for Arithmetic Fine-Tuning of Small Language Modelscited by 0