‘newest links’ directory
- See Also
- Gwern
- “Hug Dealer”, Gwern & Pro 2026
- “Human Perception at a Red Light”, Gwern & Pro 2026
- “Manual of Style”, Gwern 2025
- “Apollonian #2: A Poet Teaches AIs”, Gwern et al 2026
- “Towards a Better Hutter Prize”, Gwern 2026
- “Internet Search Case Studies”, Gwern 2019
- “The Difficulty of Psychometrics”, Gwern 2026
- “Fire Lives In The Tails”, Gwern 2026
- “Spoilage”, Pro et al 2026
- “Self-Experiment Risk-Taking Interview”, Nostrand & Gwern 2026
- “What Frightens Us In A Madman Is His Sanity”, Gwern & Pro 2026
- Links
- “Font Explorer [Gwern.net Variant]”, Stiehl & Claude-4.6-opus 2026
- “Implicit Meta-Learning May Lead Language Models to Trust More Reliable Sources”, Krasheninnikov et al 2023
- “Strategic Navigation or Stochastic Search? How Agents and Humans Reason Over Document Collections”, Borchmann et al 2026
- “BALROG”
- “Krasheninnikov/internalization: Code for the Paper ‘Implicit Meta-Learning May Lead Language Models to Trust More Reliable Sources’”
- “Optimization Lessons from a Minecraft Structure Locator”
- “Personal Encyclopedias”, Jeremy 2026
- “The Pitt’s Shabana Azeez Wants to Be the Next Robert Pattinson”
- “The Terrarium”
- “Chinese Start-Up Moonshot AI Raises US$1 Billion in Funding round Led by Alibaba and VC HongShan amid Strong Interest for OpenAI-Type Firms”
- “Marriage Over, €100,000 down the Drain: the AI Users Whose Lives Were Wrecked by Delusion”
- “ChatGPT—Human Perception at a Red Light”
- “Integer Programming Easily Encloses Horse”, Dynomight 2026
- “Apple Randomly Closes Bug Reports Unless You ‘Verify’ the Bug Remains Unfixed”
- “‘Artificial Neurons Can Be Trained to Kill Humans without Humans Overseeing’”
- “How to Be Less Awkward”, Mastroianni 2026
- “American Diner Gothic”
- “The Homogenizing Effect of Large Language Models on Human Expression and Thought”, Sourati et al 2026
- “Newtonianism and Information Control in Rome at the Wake of the 18th Century”, Macuglia 2020
- “202603”
- “Pre-Training under Infinite Compute”, Kim et al 2025
- “Scaling Karpathy’s Autoresearch: What Happens When the Agent Gets a GPU Cluster”
- “LLM Neuroanatomy II: Modern LLM Hacking and Hints of a Universal Language?”
- “Quickstrom: Property Based Acceptance Testing With LTL Specifications”, O’Connor & Wickström 2022
- “LLM in a Flash: Efficient Large Language Model Inference With Limited Memory”, Alizadeh et al 2023
- “General Exploratory Bonus for Optimistic Exploration in RLHF”, Li et al 2025
- “The Hollow Men”, Eliot 1925
- “In Search Of Banksy: The British Street Artist’s Identity Has Been Debated, and Closely Guarded, for Decades. A Quest to Solve the Riddle Took Reuters from a Bombed-Out Ukrainian Village to London and Downtown Manhattan—And Uncovered Much More Than a Name”, Gardner et al 2026
- “[4chan /vg/ Aidg Thread Screenshot of an Early Inner-Monologue for Arithmetic, Using a Katawa Shoujo Lilly Scenario]”, Anonymous 2020
- “Flash-KMeans: Fast and Memory-Efficient Exact K-Means”, Yang et al 2026
- “Warranty Void If Regenerated”, Claude-4.6-opus 2026
- “The World Is Too Much With Us”, Chatti & Pro 2026
- “Every Layer of Review Makes You 10× Slower”, apenwarr 2026
- “Characterizing Delusional Spirals through Human-LLM Chat Logs”, Moore et al 2026
- “A Purpose-Built Open Source Liquid Handler for Industry-Class Automated Experiments”, Golas et al 2026
- “Men’s Physical Strength Is Associated With Women’s Perceptions of Their Dancing Ability”, Hugill et al 2009
- “The Lean 4 Theorem Prover and Programming Language”, Moura & Ullrich 2021
- “Implicit Regularization in Deep Matrix Factorization”, Arora et al 2019
- “Mathematical Models of Computation in Superposition”, Hänni et al 2024
- “Relationship between Owner Anxiety and Feline Stress during Veterinary Consultations”, Franck et al 2026
- “On the Complexity of Random Satisfiability Problems With Planted Solutions”, Feldman et al 2013
- “Should You Marginalize over Possible Tokenizations?”, Chirkova et al 2023
- “Annotated Hands for Generative Models”, Yang et al 2024
- “Tokenization Multiplicity Leads to Arbitrary Price Variation in LLM-As-A-Service”, Chatzi et al 2025
- “Doc-To-LoRA (D2L): Learning to Instantly Internalize Contexts”, Charakorn et al 2026
- “Towards Human Computable Passwords”, Blocki et al 2014
- “OpenZL: A Graph-Based Model for Compression”, Collet et al 2025
- “Impossible by Degrees: Cohomology & Bistable Visual Paradox”, Ghrist & Ghrist 2026
- “Censored LLMs As a Natural Testbed for Secret Knowledge Elicitation”, Casademunt et al 2026
- “Elon Musk Pushes out More XAI Founders As AI Coding Effort Falters: Tesla and SpaceX Managers Sent in to Review Work As Billionaire’s Start-Up Struggles to Keep pace With Rivals”, Morris & Criddle 2026
- “How Learning to Read Changes the Cortical Networks for Vision and Language”, Dehaene et al 2010
- “Great, Now Write an Article About That: The Crescendo Multi-Turn LLM Jailbreak Attack”, Russinovich et al 2024
- “AI Use in American Newspapers Is Widespread, Uneven, and Rarely Disclosed”, Russell et al 2025
- “Lean4Physics: Comprehensive Reasoning Framework for College-Level Physics in Lean4”, Li et al 2025
- “Formalizing the Stability of the Two Higgs Doublet Model Potential into Lean: Identifying an Error in the Literature”, Tooby-Smith 2026
- “Dear Aliens: a Writing Contest for Humans [$2.5k Prizes]”, Troesh 2026
- “Learning Fast Approximations of Sparse Coding”, Gregor & LeCun 2010
- “How to Train the Best Embedding Model in the World: One PhD Later, I’m Giving My Secrets Away for Free”, Morris 2026
- “Learnings from Paying Artists Royalties for AI-Generated Art: A Retrospective on Tess.Design, Our Attempt to Make an Ethical, Artist-Friendly AI Marketplace. We Launched Tess in May 2024 and Shut It down in January 2026”, Enthoven 2026
- “Evaluating
AGENTS.md: Are Repository-Level Context Files Helpful for Coding Agents?”, Gloaguen et al 2026 - “LLM-Optimized Research Articles”, Riedel 2026
- “Neural Networks Are Decision Trees”, Aytekin 2022
- “BOLD: Boolean Logic Deep Learning”, Nguyen et al 2024
- “From Blind Spots to Gains: Diagnostic-Driven Iterative Training for Large Multimodal Models”, Jia et al 2026
- “The Clarke Tax As a Consensus Mechanism among Automated Agents”, Ephrati & Rosenschein 1991
- “Sex and Searching For Children Among Aka Foragers and Ngandu Farmers of Central Africa”, Hewlett & Hewlett 2010
- “When I Heard the Learn’d Astronomer”, Whitman 1865
- “Nudging: Inference-Time Alignment of LLMs via Guided Decoding”, Fei et al 2024
- “An Innovative New Bonus Auction for Warrant Officers”, Justicz 2026
- Wikipedia (24)
- Miscellaneous
- Bibliography
See Also
Gwern
“Hug Dealer”, Gwern & Pro 2026
“Human Perception at a Red Light”, Gwern & Pro 2026
“Manual of Style”, Gwern 2025
“Apollonian #2: A Poet Teaches AIs”, Gwern et al 2026
“Towards a Better Hutter Prize”, Gwern 2026
“Internet Search Case Studies”, Gwern 2019
“The Difficulty of Psychometrics”, Gwern 2026
“Fire Lives In The Tails”, Gwern 2026
“Spoilage”, Pro et al 2026
“Self-Experiment Risk-Taking Interview”, Nostrand & Gwern 2026
“What Frightens Us In A Madman Is His Sanity”, Gwern & Pro 2026
Links
“Font Explorer [Gwern.net Variant]”, Stiehl & Claude-4.6-opus 2026
“Implicit Meta-Learning May Lead Language Models to Trust More Reliable Sources”, Krasheninnikov et al 2023
Implicit meta-learning may lead language models to trust more reliable sources
“Strategic Navigation or Stochastic Search? How Agents and Humans Reason Over Document Collections”, Borchmann et al 2026
Strategic Navigation or Stochastic Search? How Agents and Humans Reason Over Document Collections
“BALROG”
“Krasheninnikov/internalization: Code for the Paper ‘Implicit Meta-Learning May Lead Language Models to Trust More Reliable Sources’”
“Optimization Lessons from a Minecraft Structure Locator”
“Personal Encyclopedias”, Jeremy 2026
“The Pitt’s Shabana Azeez Wants to Be the Next Robert Pattinson”
The Pitt’s Shabana Azeez Wants to Be the Next Robert Pattinson
“The Terrarium”
“Chinese Start-Up Moonshot AI Raises US$1 Billion in Funding round Led by Alibaba and VC HongShan amid Strong Interest for OpenAI-Type Firms”
“Marriage Over, €100,000 down the Drain: the AI Users Whose Lives Were Wrecked by Delusion”
Marriage over, €100,000 down the drain: the AI users whose lives were wrecked by delusion
“ChatGPT—Human Perception at a Red Light”
“Integer Programming Easily Encloses Horse”, Dynomight 2026
“Apple Randomly Closes Bug Reports Unless You ‘Verify’ the Bug Remains Unfixed”
Apple randomly closes bug reports unless you ‘verify’ the bug remains unfixed
“‘Artificial Neurons Can Be Trained to Kill Humans without Humans Overseeing’”
‘Artificial neurons can be trained to kill humans without humans overseeing’
“How to Be Less Awkward”, Mastroianni 2026
“American Diner Gothic”
“The Homogenizing Effect of Large Language Models on Human Expression and Thought”, Sourati et al 2026
The homogenizing effect of large language models on human expression and thought
“Newtonianism and Information Control in Rome at the Wake of the 18th Century”, Macuglia 2020
Newtonianism and information control in Rome at the wake of the 18th century
“202603”
View External Link:
“Pre-Training under Infinite Compute”, Kim et al 2025
“Scaling Karpathy’s Autoresearch: What Happens When the Agent Gets a GPU Cluster”
Scaling Karpathy’s Autoresearch: What Happens When the Agent Gets a GPU Cluster
“LLM Neuroanatomy II: Modern LLM Hacking and Hints of a Universal Language?”
LLM Neuroanatomy II: Modern LLM Hacking and hints of a Universal Language?
View External Link:
https://dnhkng.github.io/posts/rys-ii/#the-heatmaps-results-first
“Quickstrom: Property Based Acceptance Testing With LTL Specifications”, O’Connor & Wickström 2022
Quickstrom: Property Based Acceptance Testing with LTL Specifications
“LLM in a Flash: Efficient Large Language Model Inference With Limited Memory”, Alizadeh et al 2023
LLM in a flash: Efficient Large Language Model Inference with Limited Memory
“General Exploratory Bonus for Optimistic Exploration in RLHF”, Li et al 2025
General Exploratory Bonus for Optimistic Exploration in RLHF
“The Hollow Men”, Eliot 1925
“In Search Of Banksy: The British Street Artist’s Identity Has Been Debated, and Closely Guarded, for Decades. A Quest to Solve the Riddle Took Reuters from a Bombed-Out Ukrainian Village to London and Downtown Manhattan—And Uncovered Much More Than a Name”, Gardner et al 2026
“[4chan /vg/ Aidg Thread Screenshot of an Early Inner-Monologue for Arithmetic, Using a Katawa Shoujo Lilly Scenario]”, Anonymous 2020
![[4chan /vg/ aidg thread screenshot of an early inner-monologue for arithmetic, using a Katawa Shoujo Lilly scenario]](/doc/ai/nn/transformer/gpt/inner-monologue/2020-07-16-4chan-vg-aidg-anonymous-aidungeon2-katawashoujolilly-earlyinnermonologuepromptexample-1594872803390.png){kind=link}
“Flash-KMeans: Fast and Memory-Efficient Exact K-Means”, Yang et al 2026
“Warranty Void If Regenerated”, Claude-4.6-opus 2026
“The World Is Too Much With Us”, Chatti & Pro 2026
“Every Layer of Review Makes You 10× Slower”, apenwarr 2026
“Characterizing Delusional Spirals through Human-LLM Chat Logs”, Moore et al 2026
Characterizing Delusional Spirals through Human-LLM Chat Logs
“A Purpose-Built Open Source Liquid Handler for Industry-Class Automated Experiments”, Golas et al 2026
A Purpose-Built Open Source Liquid Handler for Industry-Class Automated Experiments
“Men’s Physical Strength Is Associated With Women’s Perceptions of Their Dancing Ability”, Hugill et al 2009
Men’s physical strength is associated with women’s perceptions of their dancing ability
“The Lean 4 Theorem Prover and Programming Language”, Moura & Ullrich 2021
“Implicit Regularization in Deep Matrix Factorization”, Arora et al 2019
“Mathematical Models of Computation in Superposition”, Hänni et al 2024
“Relationship between Owner Anxiety and Feline Stress during Veterinary Consultations”, Franck et al 2026
Relationship between owner anxiety and feline stress during veterinary consultations
“On the Complexity of Random Satisfiability Problems With Planted Solutions”, Feldman et al 2013
On the Complexity of Random Satisfiability Problems with Planted Solutions
“Should You Marginalize over Possible Tokenizations?”, Chirkova et al 2023
“Annotated Hands for Generative Models”, Yang et al 2024
“Tokenization Multiplicity Leads to Arbitrary Price Variation in LLM-As-A-Service”, Chatzi et al 2025
Tokenization Multiplicity Leads to Arbitrary Price Variation in LLM-as-a-service
“Doc-To-LoRA (D2L): Learning to Instantly Internalize Contexts”, Charakorn et al 2026
Doc-to-LoRA (D2L): Learning to Instantly Internalize Contexts
“Towards Human Computable Passwords”, Blocki et al 2014
“OpenZL: A Graph-Based Model for Compression”, Collet et al 2025
“Impossible by Degrees: Cohomology & Bistable Visual Paradox”, Ghrist & Ghrist 2026
“Censored LLMs As a Natural Testbed for Secret Knowledge Elicitation”, Casademunt et al 2026
Censored LLMs as a Natural Testbed for Secret Knowledge Elicitation
“Elon Musk Pushes out More XAI Founders As AI Coding Effort Falters: Tesla and SpaceX Managers Sent in to Review Work As Billionaire’s Start-Up Struggles to Keep pace With Rivals”, Morris & Criddle 2026
“How Learning to Read Changes the Cortical Networks for Vision and Language”, Dehaene et al 2010
How Learning to Read Changes the Cortical Networks for Vision and Language
“Great, Now Write an Article About That: The Crescendo Multi-Turn LLM Jailbreak Attack”, Russinovich et al 2024
Great, Now Write an Article About That: The Crescendo Multi-Turn LLM Jailbreak Attack
“AI Use in American Newspapers Is Widespread, Uneven, and Rarely Disclosed”, Russell et al 2025
AI use in American newspapers is widespread, uneven, and rarely disclosed
“Lean4Physics: Comprehensive Reasoning Framework for College-Level Physics in Lean4”, Li et al 2025
Lean4Physics: Comprehensive Reasoning Framework for College-level Physics in Lean4
“Formalizing the Stability of the Two Higgs Doublet Model Potential into Lean: Identifying an Error in the Literature”, Tooby-Smith 2026
“Dear Aliens: a Writing Contest for Humans [$2.5k Prizes]”, Troesh 2026
“Learning Fast Approximations of Sparse Coding”, Gregor & LeCun 2010
“How to Train the Best Embedding Model in the World: One PhD Later, I’m Giving My Secrets Away for Free”, Morris 2026
“Learnings from Paying Artists Royalties for AI-Generated Art: A Retrospective on Tess.Design, Our Attempt to Make an Ethical, Artist-Friendly AI Marketplace. We Launched Tess in May 2024 and Shut It down in January 2026”, Enthoven 2026
“Evaluating AGENTS.md: Are Repository-Level Context Files Helpful for Coding Agents?”, Gloaguen et al 2026
Evaluating AGENTS.md: Are Repository-Level Context Files Helpful for Coding Agents?
“LLM-Optimized Research Articles”, Riedel 2026
“Neural Networks Are Decision Trees”, Aytekin 2022
“BOLD: Boolean Logic Deep Learning”, Nguyen et al 2024
“From Blind Spots to Gains: Diagnostic-Driven Iterative Training for Large Multimodal Models”, Jia et al 2026
From Blind Spots to Gains: Diagnostic-Driven Iterative Training for Large Multimodal Models
“The Clarke Tax As a Consensus Mechanism among Automated Agents”, Ephrati & Rosenschein 1991
The Clarke tax as a consensus mechanism among automated agents
“Sex and Searching For Children Among Aka Foragers and Ngandu Farmers of Central Africa”, Hewlett & Hewlett 2010
Sex and Searching For Children Among Aka Foragers and Ngandu Farmers of Central Africa
“When I Heard the Learn’d Astronomer”, Whitman 1865
“Nudging: Inference-Time Alignment of LLMs via Guided Decoding”, Fei et al 2024
Nudging: Inference-time Alignment of LLMs via Guided Decoding
“An Innovative New Bonus Auction for Warrant Officers”, Justicz 2026
Wikipedia (24)
Miscellaneous
Bibliography
https://arxiv.org/abs/2310.15047: “Implicit Meta-Learning May Lead Language Models to Trust More Reliable Sources”,https://poets.org/poem/world-too-much-us-0: “The World Is Too Much With Us”,https://apenwarr.ca/log/20260316: “Every Layer of Review Makes You 10× Slower”,