AI understanding - Revision history

KevinYager: /* Psychology */

2026-04-06T21:18:48Z

‎Psychology

2026-04-02T17:48:25Z

‎Psychology

2026-04-02T17:45:07Z

‎Psychology

2026-03-18T17:37:50Z

‎Psychology

2026-03-18T12:52:09Z

‎Failure Modes

2026-03-06T22:19:50Z

‎Feature Geometry Reproduces Problem-space

2026-03-06T22:17:37Z

‎Geometric

2026-03-04T16:38:38Z

‎Geometric

2026-03-04T16:36:47Z

‎Grokking

2026-02-24T13:06:23Z

‎Psychology

@@ Line 375: / Line 375: @@
 * 2025-05: [https://arxiv.org/abs/2505.17117 From Tokens to Thoughts: How LLMs and Humans Trade Compression for Meaning]
 * 2025-07: [https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5357179 Call Me A Jerk: Persuading AI to Comply with Objectionable Requests]
 * 2026-01: [https://arxiv.org/abs/2601.06047 "They parted illusions -- they parted disclaim marinade": Misalignment as structural fidelity in LLMs]
 * 2026-02: [https://arxiv.org/abs/2602.02606 Gender Dynamics and Homophily in a Social Network of LLM Agents]

@@ Line 379: / Line 379: @@
 * 2026-02: [https://arxiv.org/abs/2602.01689 What LLMs Think When You Don't Tell Them What to Think About?]
 * 2026-03: [https://truthful.ai/consciousness_cluster.pdf The Consciousness Cluster: Preferences of Models that Claim to be Conscious]
-* 2026-04: [https://www.anthropic.com/research/emotion-concepts-function Emotion concepts and their function in a large language model]
+* 2026-04: [https://transformer-circuits.pub/2026/emotions/index.html Emotion concepts and their function in a large language model] ([https://www.anthropic.com/research/emotion-concepts-function blog])
 ==Persona Simulator Theory==

@@ Line 379: / Line 379: @@
 * 2026-02: [https://arxiv.org/abs/2602.01689 What LLMs Think When You Don't Tell Them What to Think About?]
 * 2026-03: [https://truthful.ai/consciousness_cluster.pdf The Consciousness Cluster: Preferences of Models that Claim to be Conscious]
 ==Persona Simulator Theory==

@@ Line 378: / Line 378: @@
 * 2026-02: [https://arxiv.org/abs/2602.02606 Gender Dynamics and Homophily in a Social Network of LLM Agents]
 * 2026-02: [https://arxiv.org/abs/2602.01689 What LLMs Think When You Don't Tell Them What to Think About?]
 ==Persona Simulator Theory==

@@ Line 327: / Line 327: @@
 * 2023-09: [https://arxiv.org/abs/2309.13638 Embers of Autoregression: Understanding Large Language Models Through the Problem They are Trained to Solve] (biases towards "common" numbers, in-context CoT can reduce performance by incorrectly priming, etc.)
 * 2023-11: [https://arxiv.org/abs/2311.16093 Visual cognition in multimodal large language models] (models lack human-like visual understanding)
 ==Fracture Representation==

@@ Line 142: / Line 142: @@
 * [https://arxiv.org/abs/2501.00070 ICLR: In-Context Learning of Representations]
 * [https://arxiv.org/abs/2502.00873 Language Models Use Trigonometry to Do Addition]: Numbers arranged in helix to enable addition
 ===Capturing Physics===

@@ Line 92: / Line 92: @@
 * 2025-10: [https://transformer-circuits.pub/2025/linebreaks/index.html When Models Manipulate Manifolds: The Geometry of a Counting Task]
 * 2025-10: [https://arxiv.org/abs/2510.26745 Deep sequence models tend to memorize geometrically; it is unclear why]
 * 2026-03: [https://arxiv.org/abs/2603.01968 Intrinsic Task Symmetry Drives Generalization in Algorithmic Tasks]

@@ Line 183: / Line 183: @@
 * 2024-12: [https://arxiv.org/abs/2412.09810 The Complexity Dynamics of Grokking]
 * 2025-09: [https://arxiv.org/abs/2509.21519 Provable Scaling Laws of Feature Emergence from Learning Dynamics of Grokking]
 ===Tests of Resilience to Dropouts/etc.===