{"total":415,"papers":[{"id":"arxiv:2509.03646","title":"Emergent Hierarchical Reasoning in LLMs through Reinforcement Learning","date":"2025-09-03","section":"Post-training","quality":-0.042926,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/683657af0d699b7d854ce84d95bcd53c12e152aa738c602d0982cc0d11009f08/","source_url":"https://arxiv.org/abs/2509.03646","text_available":true},{"id":"url:transformer-circuits.pub/2025/introspection","title":"Emergent Introspective Awareness in Large Language Models (Anthropic, October 2025)","date":"2025-10-29","section":"AI safety and consciousness","quality":null,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/30fa339b938a47b599e7cd074489162aa385bdb2e2d7a9a6f134d3a3a1bb5487/","source_url":"https://transformer-circuits.pub/2025/introspection/index.html","text_available":false},{"id":"arxiv:2410.07041","title":"Emergent properties with repeated examples","date":"2024-10-09","section":"Data, training, optimization","quality":0.087466,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/63a5d833b43fae5ed1f9f44f26b90813fa8a37dd44870e8e409d7a7c34fabfec/","source_url":"https://arxiv.org/abs/2410.07041","text_available":true},{"id":"arxiv:2104.14294","title":"Emerging Properties in Self-Supervised Vision Transformers","date":"2021-04-29","section":"Self-supervised learning and vision","quality":0.474747,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/4372186d141ca514b08e93bb3b1e26fdbcba9ce62187fb772c14bf223160c06d/","source_url":"https://arxiv.org/abs/2104.14294","text_available":true},{"id":"arxiv:2309.15807","title":"Emu: Enhancing Image Generation Models Using Photogenic Needles in a Haystack","date":"2023-09-27","section":"Self-supervised learning and vision","quality":0.187565,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/f1b51d5d6f9eb66c9b6424518136746ec684fa58752ed4e2ec2a7031df9fe2ad/","source_url":"https://arxiv.org/abs/2309.15807","text_available":true},{"id":"arxiv:2511.09149","title":"Enabling Agents to Communicate Entirely in Latent Space","date":"2025-11-12","section":"LLMs: architectures, context, training","quality":0.297245,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/02979751e22005b3284e4d7caa19baae335ca63c9ec46e27d833e95d088e9688/","source_url":"https://arxiv.org/abs/2511.09149","text_available":true},{"id":"arxiv:2504.06225","title":"Encoder-Decoder Gemma: Improving the Quality-Efficiency Trade-Off via Adaptation","date":"2025-04-08","section":"LLMs: architectures, context, training","quality":-0.004879,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/797af53c1331a8303ff731ef2763d562b29b9864407468ebbd7c1a89ab7c0eed/","source_url":"https://arxiv.org/abs/2504.06225","text_available":true},{"id":"arxiv:2510.26622","title":"Encoder-Decoder or Decoder-Only? Revisiting Encoder-Decoder Large Language Model","date":"2025-10-30","section":"LLMs: architectures, context, training","quality":0.064707,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/dc73f87f798d016e2d83acd1d362b4b3525af826e68b055bb5196f120d23e080/","source_url":"https://arxiv.org/abs/2510.26622","text_available":true},{"id":"arxiv:2302.07253","title":"Energy Transformer","date":"2023-02-14","section":"LLMs: architectures, context, training","quality":-0.291157,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/4f82adc1eb9a92ea3160324c195688d3834a4c68c026ffc40e951352c572746b/","source_url":"https://arxiv.org/abs/2302.07253","text_available":true},{"id":"arxiv:2605.21488","title":"Equilibrium Reasoners: Learning Attractors Enables Scalable Reasoning","date":"2026-05-20","section":"Reasoning and the \"physics\" of language models","quality":null,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/48b3beca0f016a3683df4c6c71fc4be9f13a2c12c61d4ad17d1117c6139bb11e/","source_url":"https://arxiv.org/abs/2605.21488","text_available":true},{"id":"arxiv:2512.00499","title":"ESPO: Entropy Importance Sampling Policy Optimization","date":"2025-11-29","section":"Post-training","quality":0.192562,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/342b84df2bf40f64c57ea7839579c1f337374ad9c28b16ce7d477c42e24c67b4/","source_url":"https://arxiv.org/abs/2512.00499","text_available":true},{"id":"arxiv:2502.15104","title":"Estimating Neural Representation Alignment from Sparsely Sampled Inputs and Features","date":"2025-02-20","section":"Representation alignment","quality":0.498999,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/bb284056f9389cbe75ad82bc3d85b997ce69212d23f95de2baf599ad64d97618/","source_url":"https://arxiv.org/abs/2502.15104","text_available":true},{"id":"arxiv:2406.03689","title":"Evaluating the World Model Implicit in a Generative Model","date":"2024-06-06","section":"Reasoning and the \"physics\" of language models","quality":0.186581,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/41da4da8a33d9ab26af9650748067540eaf4cd386071ae0224a06450cdbb1fd0/","source_url":"https://arxiv.org/abs/2406.03689","text_available":true},{"id":"doi:10.1073/pnas.2520095123","title":"Evidence from formal logical reasoning reveals that the language of thought is not natural language","date":"2026-07-06","section":"Reasoning and the \"physics\" of language models","quality":-0.070856,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/c9ba9811c4b4f1ea1fa65751f033bad4c2f6543f454b4d0d9e91a52abc6eb407/","source_url":"https://doi.org/10.1073/pnas.2520095123","text_available":true},{"id":"arxiv:2601.20861","title":"Evolutionary Strategies lead to Catastrophic Forgetting in LLMs","date":"2026-01-28","section":"Post-training","quality":-0.530464,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/4178758b0cb3ebc95f6a3cbc6423e8ea05c0ae649e11a2f4059a84b3faa6cde0/","source_url":"https://arxiv.org/abs/2601.20861","text_available":true},{"id":"arxiv:2607.27372","title":"Explorative Modeling: Unlocking a Third Pretraining Axis and End-to-End Generation","date":"2026-07-29","section":"Data, training, optimization","quality":0.296593,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/542a5e85b84624381953f643ec6075a588c08c542c4f945e57d39c9164c9a835/","source_url":"https://arxiv.org/abs/2607.27372","text_available":true},{"id":"arxiv:2604.08224","title":"Externalization in LLM Agents: A Unified Review of Memory, Skills, Protocols and Harness Engineering","date":"2026-04-09","section":"Harness","quality":-0.194929,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/392317a8b3bedec8258ff8288d69ac72bc9e60299c5f7bec8962838bb448c2ce/","source_url":"https://arxiv.org/abs/2604.08224","text_available":true},{"id":"arxiv:1807.02787","title":"Financial Trading as a Game: A Deep Reinforcement Learning Approach","date":"2018-07-08","section":"Finance","quality":-0.745254,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/ffab71fb6c4e10245324b584ff3252a6fae8005a8220f84f1c44da083caec855/","source_url":"https://arxiv.org/abs/1807.02787","text_available":true},{"id":"arxiv:2609.16454","title":"Fine-Tuning Fixes Mode Collapse and Over-Dispersion in LLMs","date":"2026-09-15","section":"Post-training","quality":null,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/96b9ce3c41bf5be0185a59605d2cade1d5dcdc9432040a1195506fac45cb96fa/","source_url":"https://arxiv.org/abs/2609.16454","text_available":true},{"id":"arxiv:2610.02140","title":"Finetuning with Sampling: SFT Learns Better Than You Think","date":"2026-10-01","section":"Post-training","quality":null,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/60f5edfae92f505bf86e8890135a44b4129825b0525b5ad8d242e8ba04182605/","source_url":"https://arxiv.org/abs/2610.02140","text_available":true},{"id":"url:weco.ai/blog/first-evidence-of-recursive-self-improvement","title":"First evidence of recursive self-improvement (AIDE2, Weco)","date":"2026-07-14","section":"Harness","quality":null,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/00441fa8f2a9743603579837e07f96b97a0a7de239cad14faaccc9aa2e9d770c/","source_url":"https://www.weco.ai/blog/first-evidence-of-recursive-self-improvement","text_available":false},{"id":"arxiv:2004.12919","title":"First return, then explore","date":"2020-04-27","section":"Reinforcement learning","quality":0.582832,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/3cd790c501eaa633193918533947c58e30d1939bc08577606f1d88c56d368285/","source_url":"https://arxiv.org/abs/2004.12919","text_available":true},{"id":"arxiv:2606.18206","title":"Fixed-Point Reasoners: Stable and Adaptive Deep Looped Transformers","date":"2026-06-16","section":"Reasoning and the \"physics\" of language models","quality":null,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/37ea2d83a92ece4ab3ba4434af71f291d1fdc5040676f75110210f40aad3b81a/","source_url":"https://arxiv.org/abs/2606.18206","text_available":true},{"id":"arxiv:2311.06242","title":"Florence-2: Advancing a Unified Representation for a Variety of Vision Tasks","date":"2023-11-10","section":"LLMs: architectures, context, training","quality":0.228032,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/edc2679fff98a10c70395e79c58b56eaf9ec0001691dbca2b33ba94a401a8398/","source_url":"https://arxiv.org/abs/2311.06242","text_available":true}],"next":"https://vinnik-dmitry07.github.io/favorite-papers/agents/catalog-0-6.html"}