AI Wiki — Knowledge Base

图灵奖获得者、 AI 三大教父之一的 LeCun在达沃斯的发言，算是把整个硅谷的遮羞布扯了。他说现在整个行业都被LLM彻底洗脑了，所有人都在同一条赛道上互相挖人，谁敢偏离主流谁就被骂落后。这也是他离开Meta的真正原因，连Meta都已经LLM-pilled，他不想再跟风了。 …

图灵奖获得者、 AI 三大教父之一的 LeCun在达沃斯的发言，算是把整个硅谷的遮羞布扯了。他说现在整个行业都被LLM彻底洗脑了，所有人都在同一条赛道上互相挖人，谁敢偏离主流谁就被骂落后。这也是他离开Meta的真正原因，连Meta都已经LLM-pilled，他不想再跟风了。最扎心的一句话是：纯生成式架构，不管是LLM、VLM还是VLA，永远造不出哪怕猫级的智能体。因为它们本质上只是下一个t

researchyann-lecun--x

24 Apr 2026

Search: “llm”

AP-BMM: Approximating Capability-Efficiency Pareto Sets of LLMs via Asynchronous Prior-guided Bayesian Model Merging

CFDLLMBench: A Benchmark Suite for Evaluating Large Language Models in Computational Fluid Dynamics

Guided Speculative Inference for Efficient Test-Time Alignment of LLMs

Is Chain-of-Thought Reasoning of LLMs a Mirage? A Data Distribution Lens

Self Knowledge Re-expression: A Fully Local Method for Adapting LLMs to Tasks Using Intrinsic Knowledge

Thoughtful book on LLMs and linguistics — by someone who actually knows linguistics.

Utility-Aware Data Pricing: Token-Level Quality and Empirical Training Gain for LLMs

Scaling massive monolithic LLMs continues to yield incredible results. But to truly unlock their ceiling, the next frontier is test-time com…

Context Is What You Need: The Maximum Effective Context Window for Real World Limits of LLMs

Leveraging Multimodal LLMs for Built Environment and Housing Attribute Assessment from Street-View Imagery

Not-a-Bandit: Provably No-Regret Drafter Selection in Speculative Decoding for LLMs

Wiring the 'Why': A Unified Taxonomy and Survey of Abductive Reasoning in LLMs

Can LLMs Infer Conversational Agent Users' Personality Traits from Chat History?

Caught in the Web of Words: Do LLMs Fall for Spin in Medical Literature?

Auditing LLMs for Algorithmic Fairness in Casenote-Augmented Tabular Prediction

Beyond Marginal Distributions: A Framework to Evaluate the Representativeness of Demographic-Aligned LLMs

Cyber Defense Benchmark: Agentic Threat Hunting Evaluation for LLMs in SecOps

Decomposed Trust: Privacy, Adversarial Robustness, Ethics, and Fairness in Low-Rank LLMs

FedProxy: Federated Fine-Tuning of LLMs via Proxy SLMs and Heterogeneity-Aware Fusion

LBLLM: Lightweight Binarization of Large Language Models via Three-Stage Distillation

CodePivot: Bootstrapping Multilingual Transpilation in LLMs via Reinforcement Learning without Parallel Corpora

CoLLM: A Unified Framework for Co-execution of LLMs Federated Fine-tuning and Inference

Counterfactual Modeling with Fine-Tuned LLMs for Health Intervention Design and Sensor Data Augmentation

Emergent Misalignment via In-Context Learning: Narrow in-context examples can produce broadly misaligned LLMs

FedLLM: A Privacy-Preserving Federated Large Language Model for Explainable Traffic Flow Prediction

GSQ: Highly-Accurate Low-Precision Scalar Quantization for LLMs via Gumbel-Softmax Sampling

LLMs can persuade only psychologically susceptible humans on societal issues, via trust in AI and emotional appeals, amid logical fallacies

ResearchBench: Benchmarking LLMs in Scientific Discovery via Inspiration-Based Task Decomposition

The Potential of Second-Order Optimization for LLMs: A Study with Full Gauss-Newton

Pruning Unsafe Tickets: A Resource-Efficient Framework for Safer and More Robust LLMs

Self-Distillation as a Performance Recovery Mechanism for LLMs: Counteracting Compression and Catastrophic Forgetting

Structured Abductive-Deductive-Inductive Reasoning for LLMs via Algebraic Invariants

What Makes LLMs Effective Sequential Recommenders? A Study on Preference Intensity and Temporal Context

WildFeedback: Aligning LLMs With In-situ User Interactions And Feedback

DiscoTrace: Representing and Comparing Answering Strategies of Humans and LLMs in Information-Seeking Question Answering

Language on Demand, Knowledge at Core: Composing LLMs with Encoder-Decoder Translation Models for Extensible Multilinguality

RL-PLUS: Countering Capability Boundary Collapse of LLMs in Reinforcement Learning with Hybrid-policy Optimization

AgenticAI-DialogGen: Topic-Guided Conversation Generation for Fine-Tuning and Evaluating Short- and Long-Term Memories of LLMs

Calibration-Aware Policy Optimization for Reasoning LLMs

E2LLM: Encoder Elongated Large Language Models for Long-Context Understanding and Reasoning

Schema-Adaptive Tabular Representation Learning with LLMs for Generalizable Multimodal Clinical Reasoning

Aligning What LLMs Do and Say: Towards Self-Consistent Explanations

Backdoors in RLVR: Jailbreak Backdoors in LLMs From Verifiable Reward

How LLMs Might Think

Persona Non Grata: Single-Method Safety Evaluation Is Incomplete for Persona-Imbued LLMs

RL makes MLLMs see better than SFT

Single-Agent LLMs Outperform Multi-Agent Systems on Multi-Hop Reasoning Under Equal Thinking Token Budgets