Definition
Ablation studies are experiments that selectively remove, disable, or replace one component of a system to measure how much that piece contributes to overall performance. They are the workhorse method for arguing causality in ML papers: if accuracy collapses when you delete a module, that module was doing real work.
Episodes covering this
- 210Same Website Request, Different Code — The Bias You Can't SeeBiased or Personalized? The Impact of Personal Information on AI-driven Development· ·14 min·Jul 09, 2026
- 207An AI Graded Its Own Math Test 94 Percent — It Actually Scored 20More Convincing, Not More Correct: Self-Play Reward Hacking of Reference-Free LLM Judges· ·12 min·Jul 08, 2026
- 206How Four-Second Clips Become Hours of Playable AI SoccerMultiplayer Interactive World Models with Representation Autoencoders· ·15 min·Jul 07, 2026
- 204The Length Estimate Hiding Inside a Word-by-Word ModelHow Much is Left? LLMs Linearly Encode Their Remaining Output Length· ·14 min·Jul 07, 2026
- 200The One Mechanism That Turns Twenty AI Clones Into an Actual TeamEVOCHAMBER: Test-Time Co-evolution of Multi-Agent System at Individual, Team, and Population ScalesZhang, Xu, Dai et al. · Oregon State University; AG2AI·19 min·Jul 04, 2026
- 198The Model That Knows the Answer and Can't Say ItCan Language Models Actually Retrieve In-Context? Drowning in Documents at Million Token ScaleGollapudi, Gupta, Singhal et al. · UC Berkeley·17 min·Jul 03, 2026
- 197Twin Problems Suggest AI Reasoning Gains Are Mostly Better Fact RecallIsoSci: A Benchmark of Isomorphic Cross-Domain Science Problems for Evaluating Reasoning versus Knowledge Retrieval in LLMsAbdaljalil, Serpedin, Kurban · Texas A&M University·17 min·Jul 03, 2026
- 193Freeze Most of the Network: Where RL Improvement Actually Lives in a TransformerIs One Layer Enough? Training A Single Transformer Layer Can Match Full-Parameter RL TrainingZhang, Hu, Glentis et al. · University of Minnesota·22 min·Jul 02, 2026
- 182How a Tiny Model Too Weak to Plan Cuts a Bigger Agent's Hallucinations by 80%Grounded Iterative Language Planning: How Parameterized World Models Reduce Hallucination Propagation in LLM AgentsSong, Cai · Emory University·17 min·Jun 29, 2026
- 178How an AI Reviewer Learned to Stop Going Easy on AI WritingThe Red Queen Gödel Machine: Co-Evolving Agents and Their EvaluatorsIacob, Jovanović, Shen et al. · University of Cambridge·23 min·Jun 26, 2026
- 169Why Better Bug Reports Can Make AI Coding Agents WorseSHERLOC: Structured Diagnostic Localization for Code Repair AgentsTamoyan, Narenthiran, Arakelyan et al. · NVIDIA / TU Darmstadt·24 min·Jun 24, 2026
- 168When Turning Experience Into Code Makes Your AI Agent DumberMetis: Bridging Text and Code Memory for Self-Evolving AgentsDai, He, Li et al. · The Chinese University of Hong Kong·27 min·Jun 24, 2026
- 154How a 7B Model Out-Investigates a 72B One by Choosing What to Look AtNative Active Perception as Reasoning for Omni-Modal UnderstandingXing, Xu, Wang et al. · The Chinese University of Hong Kong·21 min·Jun 18, 2026
- 151Why More Experience Made This AI Agent Worse, And How to Fix ItNot All Skills Help: Measuring and Repairing Agent KnowledgeWang, Zhou, Liang et al. · UNC Chapel Hill·28 min·Jun 16, 2026
- 144When an AI Agent Just Copies Its Tool — And Bigger Models Copy MoreWhen the Tool Decides: LLM Agents Defer Blindly to Graph Neural Network Tools, and Stronger Backbones Defer MoreWang, Vemuri · raptorX.ai·15 min·Jun 15, 2026
- 139When Optimizing One GPU Kernel Quietly Breaks the Whole SystemArbor: Tree Search as a Cognition Layer for Autonomous AgentsPrakriya, Hou, Gong et al. · AMD·30 min·Jun 12, 2026
- 131Why Autonomous Research Agents Forget Their Own Lessons, and Arbor's FixToward Generalist Autonomous Research via Hypothesis-Tree RefinementJin, Hu, Qiu et al. · Renmin University of China·33 min·Jun 11, 2026
- 127What Diffusion Language Models Were Missing: A Map, Not an AlgorithmTextLDM: Language Modeling with Continuous Latent DiffusionJiang, Ren, Li et al. · JoyFuture Academy / HIT·30 min·Jun 11, 2026
- 121When the Agent Says It's Done But Nothing Happened: Debugging the Harness, Not the ModelFrom Failed Trajectories to Reliable LLM Agents: Diagnosing and Repairing Harness FlawsChen, Wang, Liu et al. · Institute of Software·27 min·Jun 05, 2026
- 120How an AI Agent Rewrites Its Own Tools, Without an Answer KeyRetrospective Harness Optimization: Improving LLM Agents via Self-Preference over Trajectory RolloutsPan, Liu, Lin et al. · City University of Hong Kong·30 min·Jun 05, 2026
- 115Teaching a Phone Agent to Reason Silently, And Keeping It HonestMIRAGE: Mobile Agents with Implicit Reasoning and Generative World ModelsYang, Hu, Hao et al. · Beihang University·24 min·Jun 04, 2026
- 111How a 4B Web Agent Beat Models 60x Its Size on 500 DemonstrationsOpenWebRL: Demystifying Online Multi-turn Reinforcement Learning for Visual Web AgentsYang, Wu, Chen et al. · UIUC·24 min·Jun 03, 2026
- 107How a Market of Crippled AI Agents Outscored One Unrestricted ModelEconomy of Minds: Emerging Multi-Agent Intelligence with Economic InteractionsQi, Su, Qu et al. · Harvard·26 min·Jun 03, 2026
- 106Giving Agents a Notebook Instead of New Weights: How ExpGraph Lets Frozen Models LearnExpGraph: Model-Agnostic Experience Learning with Graph-Structured Memory for LLM AgentsFeng, Ye, Luo et al. · University of Illinois Urbana-Champaign·26 min·Jun 02, 2026
- 105The Trojan Is Your Agent's Memory: Why Single-Step Defenses Miss Persistent AttacksFrom Prompt Injection to Persistent Control: Defending Agentic Harness Against Trojan BackdoorsTan, Dou, Yang et al. · Gaoling School of Artificial Intelligence·26 min·Jun 01, 2026
- 101Treating Math Formalization Like a Codebase, and Where the Agents CheatFormalizing Mathematics at ScaleRammal, Patel, Gloeckle et al. · FAIR at Meta / CERMICS·27 min·May 29, 2026
- 100How a Prompt Wrapper Lets a Frontier Model Play Poker Like an ExpertPokerSkill: LLMs Can Play Expert-Level Poker without Training or SolversLi, Wang, Huang · IIIS·29 min·May 29, 2026
- 097Same Tokens, Same Cost, Wildly Different Results: What Actually Scales in AI AgentsScaling Laws for Agent Harnesses via Effective Feedback ComputeZhang, Wang, Xu et al. · Harbin Institute of Technology·25 min·May 29, 2026
- 095Seven Wins to Zero: How Organizing AI Agents Like a Lab Changes the SearchAutoScientists: Self-Organizing Agent Teams for Long-Running Scientific ExperimentationGao, Fang, Zitnik · Harvard University·24 min·May 28, 2026
- 083Training the Translator: How a Small Communication Model Lets Agent Teams Outperform ThemselvesAgentFugue: Agent Scaling for Long-Horizon Tasks through Collective ReasoningHu, Qian, Wang et al. · GSAI·24 min·May 26, 2026
- 079An Old Idea From Cognitive Psychology Reshapes How We Reward Reasoning ModelsMetacognition as Reward: Reinforcing LLM Reasoning via Knowledge and Regulation SignalsChen, Xu, Zhao et al. · Tongji University / Shanghai AI Laboratory / Nanyang Technological University·29 min·May 25, 2026
- 078Training a Markdown File: When LLM Self-Improvement Borrows the Discipline of Neural Net TrainingSkillOpt: Executive Strategy for Self-Evolving Agent SkillsYang, Gong, Huang et al. · Microsoft·28 min·May 25, 2026
- 077Reading a Model's Confidence Curve to Decide When Chain-of-Thought Is Worth ItWhen Do LLMs Reason? A Dynamical Systems View via Entropy Phase TransitionsXia, Wang, Tang et al. · State Key Laboratory of General Artificial Intelligence·22 min·May 25, 2026
- 076Same Model, Organized Differently: How an Agent Architecture Beat Frontier Systems at Research MathRMA: an Agentic System for Research-Level Mathematical ProblemsZhao, Yuan, Choi et al. · Georgia Institute of Technology·22 min·May 25, 2026
- 075Growing Code and Proof Together: Verified Systems in Ten Hours Instead of a YearInductive Deductive Synthesis: Enabling AI to Generate Formally Verified SystemsAgarwal, Krentsel, Liu et al. · UC Berkeley·28 min·May 25, 2026
- 074How a Fifteen-Hundred-Dollar Training Run Matched Llama and Gemma on ReasoningHRM-Text: Efficient Pretraining Beyond ScalingWang, Liu, Wang et al. · Sapient Intelligence·21 min·May 24, 2026
- 071When the Model Is Fine and the Plumbing Is Broken: Fixing Agents at the InterfaceAdapting the Interface, Not the Model: Runtime Harness Adaptation for Deterministic LLM AgentsXu, Wen, Li · Peking University·23 min·May 22, 2026
- 070When Models Know the Answer But Say the Wrong Thing AnywayHallucination as Commitment Failure: Larger LLMs Misfire Despite Knowing the AnswerYeom, Sok, Kim et al. · Graduate School of Data Science·22 min·May 22, 2026
- 069When Smarter Models Forecast Worse: The Hidden Failure Mode in LLM PredictionsIs Capability a Liability? More Capable Language Models Make Worse Forecasts When It Matters MostMerrill, Lee, Karger · Forecasting Research Institute / UC Berkeley·30 min·May 22, 2026
- 064When Agent Memory Stops Being a Database and Starts Being a SkillAuto-Dreamer: Learning Offline Memory Consolidation for Language AgentsYe, Liu, Wang et al. · University of Illinois Urbana-Champaign·30 min·May 22, 2026
- 053An AI Agent Swapped In Focal Loss And Beat A Human-Tuned Training ScriptAgentic Discovery of Neural Architectures: AIRA-Compose and AIRA-DesignPepe, Lin, Magka et al. · FAIR at Meta·32 min·May 18, 2026
- 041When the Iteration Teaches the Model to Skip the IterationSolve the Loop: Attractor Models for Language and ReasoningFein-Ashley, Rashidinejad · University of Southern California·30 min·May 13, 2026
- 040Two Frozen Models Learn to Whisper: Coupling Through Hidden StatesThe Bicameral Model: Bidirectional Hidden-State Coupling Between Parallel Language ModelsFlamant, Ghai, Shimizu · AWS Agentic AI·29 min·May 13, 2026
- 033Echo: The Paper Arguing You Never Needed a KV Cache for RetrievalEcho: KV-Cache-Free Associative Recall with Spectral Koopman OperatorsSridhar, Johansen · California·24 min·May 11, 2026
- 021Ten Thousand Examples Beat the Full Industrial Pipeline for Search AgentsOpenSeeker-v2: Pushing the Limits of Search Agents with Informative and High-Difficulty TrajectoriesDu, Ye, Tang et al. · Shanghai Jiao Tong University·14 min·May 06, 2026
- 013Why Search Keeps Rediscovering the Same Workflow, and What That MeansWhy Search When You Can Transfer? Amortized Agentic Workflow Design from Structural PriorsDu, Liu, Du et al. · Carnegie Mellon University·22 min·May 03, 2026
- 012Why AI Coding Agents Keep Trying to Debug Without a DebuggerDynamic analysis enhances issue resolutionLiu, Wang, Chen et al. · Sun Yat-sen University·21 min·May 02, 2026