Intel
Page 35
Lilian Weng@lilianweng
Lilian Weng announces a long-planned post on scaling laws: how they let you reason about the optimal split of compute between data and model size before committing to an expensive run.
Why it mattersA practical explainer on what scaling laws predict and how to use them to allocate compute between data and model size before committing to an expensive training run — useful if you're planning or budgeting large model training.
Interconnects: Frontier Post-Training Recipe Review
A podcast interview and technical summary tracing how large language model post-training recipes have evolved from InstructGPT to 2026-era frontier models.
Why it mattersIt gives ML practitioners a clear, comparative map of how post-training recipes differ across DeepSeek, Llama, Tülü, OLMo, Nemotron, Kimi, and GLM, and explains the shift toward multi-teacher on-policy distillation.
articleNathan Lambert
How Open Model Ecosystems Compound
An analytical essay examining China's high-participation, open-first AI ecosystem and how sharing research reduces development costs across labs.
Why it mattersIt reframes the open-vs-closed model debate around R&D economics — arguing open models cost more off-the-shelf but compound advantages for teams doing internal development.
articleNathan Lambert
Notes from Inside China's AI Labs
A firsthand essay from Interconnects AI documenting a trip to visit most of China's leading AI labs, exploring the cultural, organizational.
Why it mattersA rare insider account of how China's top AI labs actually operate — their student-heavy teams, fast-follower dynamics, and 'build-not-buy' data mentality.
articleNathan Lambert
The Distillation Panic
An opinion essay from the Interconnects AI newsletter arguing against conflating legitimate model distillation with 'distillation attacks' by Chinese labs.
Why it mattersIf you build on or distill from open-weight models, this clarifies how 'distillation attacks' rhetoric could translate into regulation that restricts a workflow you likely depend on.
articleNathan Lambert
Interconnects AI: Reading Today's Open-Closed Performance Gap
An analytical essay from the Interconnects AI newsletter examining why the perceived capability gap between open and closed language models is more nuanced than a single benchmark number suggests.
Why it mattersIf you're deciding which models to build on, this reframes the open-vs-closed 'gap' as benchmark-dependent and explains why fast-follower labs stay close — useful for judging when open weights are good enough for your use case.
articleNathan Lambert
RLHF Book & Post-training Course
A comprehensive book and free lecture series teaching post-training techniques for language models, from RLHF foundations to reinforcement learning algorithms.
Why it mattersIf you're moving from prompting into actually shaping model behavior, this pairs a professionally edited RLHF book with a free structured lecture course covering post-training from foundations to modern RL algorithms.
articleNathan Lambert
Training an LLM-RecSys Hybrid for Steerable Recs with Semantic IDs
An LLM that can converse in English & item IDs, and make recommendations w/o retrieval or tools.
Why it mattersIt shows how to fold recommendation capability directly into an LLM via semantic item IDs, eliminating separate retrieval/tool infrastructure — useful for anyone building steerable, conversational recommender systems.
articleEugene Yan
An LLM-as-Judge Won't Save The Product—Fixing Your Process Will
Applying the scientific method, building via eval-driven development, and monitoring AI output.
Why it mattersIf you're relying on an LLM-as-Judge to guarantee product quality, this makes the case that the judge is only as good as your process.
articleEugene Yan
AlignEval: Building an App to Make Evals Easy, Fun, and Automated
Look at and label your data, build and evaluate your LLM-evaluator, and optimize it against your labels.
Why it mattersIf you're building LLM-as-judge evaluators, this walks through a practical loop for labeling data and optimizing your evaluator against those labels — grounding eval quality in human-aligned measurement rather than vibes.
articleEugene Yan
Evaluating the Effectiveness of LLM-Evaluators (aka LLM-as-Judge)
Use cases, techniques, alignment, finetuning, and critiques against LLM-evaluators.
Why it mattersIf you're building LLM-as-Judge evaluators, this breaks down alignment techniques, finetuning approaches, and the concrete failure modes of using LLMs to grade LLMs.
articleEugene Yan
AI Engineer 2024 Keynote - What We Learned from a Year of LLMs
Special double-feature closing keynote from the 6 authors of the hit O'Reilly article on Applied LLMs.
Why it mattersA concentrated set of production LLM lessons from six practitioners covering evals, prompting, RAG vs. fine-tuning tradeoffs, and operational pitfalls—useful if you're moving an LLM feature from demo to reliable product.
articleEugene Yan
What We've Learned From A Year of Building with LLMs
From the tactical nuts & bolts to the operational day-to-day to the long-term business strategy.
Why it mattersA dense, practitioner-written distillation of what actually works when shipping LLM apps — covering prompting patterns, RAG, evaluation strategy, and the operational/organizational realities most tutorials skip.
articleEugene Yan
Don't Mock Machine Learning Models In Unit Tests
How unit testing machine learning code differs from typical software practices
Why it mattersIf you write tests for ML pipelines, this explains why mocking models hides real failures and what to validate instead — practical guidance for keeping AI code reliable as it evolves.
articleEugene Yan
How to Generate and Use Synthetic Data for Finetuning
Overcoming the bottleneck of human annotations in instruction-tuning, preference-tuning, and pretraining.
Why it mattersIf you're finetuning models and blocked by the cost of human annotation, this walks through concrete synthetic data generation strategies for instruction-tuning, preference-tuning, and pretraining.
articleEugene Yan
Generation configurations: temperature, top-k, top-p, and test time compute
How generation is actually configured.
Why it mattersIt demystifies why LLM outputs are inconsistent and how sampling knobs (temperature, top-k, top-p) plus test-time compute actually shape responses.
articleChip Huyen
Out-of-Domain Finetuning to Bootstrap Hallucination Detection
How to use open-source, permissive-use data and collect less labeled samples for our tasks.
Why it mattersIf you're building hallucination detection or other classifiers with limited labeled data, this shows how to bootstrap from out-of-domain, permissively-licensed datasets to cut annotation costs while still hitting task performance.
articleEugene Yan
Multimodality and Large Multimodal Models (LMMs)
A primer on multimodality and large multimodal models.
Why it mattersA clear, foundational-to-frontier walkthrough of how multimodal models actually work — from CLIP and Flamingo to modern adapter-based LMMs — giving engineers the conceptual grounding to reason about and build with vision-language systems.
articleChip Huyen
Open challenges in LLM research
Ten major open research directions in language models, drawn from conversations across industry and academia.
Why it mattersA structured tour of the 10 hardest open problems in LLM research — from measuring hallucination to non-GPU compute — with curated primary papers for each, giving engineers a map of where the field is heading and what to watch.
articleChip Huyen
LLM Powered Autonomous Agents
A survey of LLM-powered autonomous agents built around planning, memory and tool use, drawing on AutoGPT, GPT-Engineer and BabyAGI as early proofs of concept.
Why it mattersA clear, canonical breakdown of the core components of LLM agents — planning with subgoal decomposition and reflection, short/long-term memory via vector stores, and tool use.
articleLilian Weng
Prompt Engineering
A survey of prompt engineering as in-context steering of an autoregressive model without touching its weights, treated as an empirical science whose methods vary sharply between models.
Why it mattersA rigorous, practitioner-oriented survey of in-context prompting methods — few-shot, chain-of-thought, self-consistency, and more.
articleLilian Weng
The Transformer Family Version 2.0
A rewritten and roughly doubled survey of transformer architecture variants, restructured and brought up to date with the improvements proposed since the original 2020 version.
Why it mattersA thorough, notation-consistent reference on transformer architecture variants — from attention mechanisms to positional encodings and efficiency tricks.
articleLilian Weng
Large Transformer Model Inference Optimization
Why inference on large transformers is hard and what actually helps.
Why it mattersA rigorous survey of how to cut transformer inference cost in time and memory — quantization, pruning, sparsity, and distillation — giving engineers a practical map for deploying large models affordably at scale.
articleLilian Weng
Generalized Visual Language Models
Vision-language systems traditionally bolt an object detector onto a text decoder.
Why it mattersA structured survey of how to graft visual understanding onto pre-trained language models, covering the design choices behind modern VLMs — useful grounding for anyone building or fine-tuning multimodal AI systems.
articleLilian Weng
Learning with not Enough Data Part 3: Data Generation
Approaches to generating synthetic training data when real data is scarce, covering augmentation of existing samples and outright generation of new ones.
Why it mattersIf you're training or fine-tuning models with limited labeled data, this breaks down practical approaches to synthesizing training examples — augmentation techniques and few-shot LM-based generation.
articleLilian Weng
Learning with not Enough Data Part 2: Active Learning
This is part 2 of what to do when facing a limited amount of labeled data for supervised learning tasks.
Why it mattersIf you're training models with scarce labels and a tight annotation budget, this walks through active learning methods for choosing which samples are worth labeling.
articleLilian Weng
Learning with not Enough Data Part 1: Semi-Supervised Learning
When facing a limited amount of labeled data for supervised learning tasks, four approaches are commonly discussed.
Why it mattersA clear walkthrough of the main approaches for training when labeled data is scarce, giving practitioners concrete semi-supervised learning strategies to squeeze more value from unlabeled datasets.
articleLilian Weng
What are Diffusion Models?
[Updated on 2021-09-19: Highly recommend this blog post on score-based generative modeling by Yang Song (author of several key papers in the references)]. [Updated on 2022-08-27.
Why it mattersA rigorous, continuously-updated technical walkthrough of diffusion models — from the DDPM math to consistency models and latent diffusion.
articleLilian Weng
Contrastive Representation Learning
The goal of contrastive representation learning is to learn such an embedding space in which similar sample pairs stay close to each other while dissimilar ones are far apart.
Why it mattersA grounding on how contrastive learning builds embedding spaces where similar samples cluster and dissimilar ones separate.
articleLilian Weng
Reducing Toxicity in Language Models
Large pretrained language models are trained over a sizable collection of online data. They unavoidably acquire certain toxic behavior and biases from the Internet.
Why it mattersIf you're deploying LLMs in production, this breaks down concrete techniques for controlling toxic and biased generation—the safety layer that separates a demo from a shippable product.
articleLilian Weng
Controllable Neural Text Generation
[Updated on 2021-02-01: Updated to version 2.0 with several work added and many typos fixed.] [Updated on 2021-05-26.
Why it mattersA deep technical survey of how to steer language model outputs — covering guided decoding, prompt tuning, and unlikelihood training — for anyone who needs finer control over what an LLM generates than raw prompting provides.
articleLilian Weng
Neural Architecture Search
Although most popular and successful model architectures are designed by human experts, it doesn’t mean we have explored the entire network architecture space and settled down with the best option.
Why it mattersExplains how model architectures can be discovered automatically rather than hand-designed by experts, covering the search space, strategy, and performance estimation that underpin NAS.
articleLilian Weng
Exploration Strategies in Deep Reinforcement Learning
Exploration versus exploitation in deep reinforcement learning.
Why it mattersA structured survey of modern exploration methods in deep RL — count-based bonuses, curiosity/forward-dynamics, and exploration via disagreement.
articleLilian Weng
Curriculum for Reinforcement Learning
[Updated on 2020-02-03: mentioning PCG in the “Task-Specific Curriculum” section. [Updated on 2020-02-04: Add a new “curriculum through distillation&rdquo.
Why it mattersA structured deep-dive into how curriculum learning accelerates and stabilizes RL training — covering task ordering, procedural content generation, and distillation-based curricula.
articleLilian Weng
Self-Supervised Representation Learning
A survey of self-supervised representation learning, covering contrastive predictive coding and the momentum-contrast family including MoCo, SimCLR and BYOL.
Why it mattersA deep, well-organized survey of self-supervised representation learning methods (CPC, MoCo, SimCLR, BYOL) that gives engineers the conceptual grounding to build and choose embedding/pretraining approaches for AI-native products.
articleLilian Weng
Evolution Strategies
Stochastic gradient descent is a universal choice for optimizing deep learning models. However, it is not the only option.
Why it mattersExplains how evolution strategies can optimize objectives where gradients are unavailable or unreliable.
articleLilian Weng
Meta Reinforcement Learning
In my earlier post on meta-learning , the problem is mainly defined in the context of few-shot classification.
Why it mattersA clear, technical walkthrough of meta-RL methods for training agents that generalize to unseen tasks quickly — useful background for anyone building adaptive or few-shot-capable RL agents.
articleLilian Weng
Domain Randomization for Sim2Real Transfer
Deep RL is too sample-hungry to train on real robots, so models are trained in simulation and fail on the reality gap.
Why it mattersIf you're training RL policies in simulation and struggling to deploy them on physical robots, this breaks down how domain randomization over physical parameters (friction, mass, damping) closes the reality gap and where naive sim modeling fails.
articleLilian Weng
Are Deep Neural Networks Dramatically Overfitted?
[Updated on 2019-05-27.
Why it mattersA rigorous walkthrough of why overparameterized deep networks generalize instead of overfitting, covering the Lottery Ticket Hypothesis, intrinsic dimension, and generalization bounds.
articleLilian Weng
Meta-Learning: Learning to Learn Fast
[Updated on 2019-10-01: thanks to Tianhao, we have this post translated in Chinese !]
Why it mattersA rigorous, well-organized primer on meta-learning that walks through metric-based, model-based, and optimization-based approaches (including MAML and memory-augmented networks) with the math and intuition.
articleLilian Weng
Flow-based Deep Generative Models
GANs and VAEs never explicitly learn the density of real data because the integral is intractable.
Why it mattersA rigorous walkthrough of normalizing flows and models like RealNVP and Glow that explicitly learn tractable data likelihoods — useful for practitioners who need exact density estimation rather than the implicit distributions of GANs/VAEs.
articleLilian Weng
From Autoencoder to Beta-VAE
A tour from the plain autoencoder through denoising, sparse and contractive variants to the variational autoencoder and beta-VAE.
Why it mattersA single, rigorous walkthrough of the autoencoder-to-VAE lineage — including VQ-VAE, VQ-VAE-2, and TD-VAE.
articleLilian Weng
Attention? Attention!
[Updated on 2018-10-28: Add Pointer Network and the link to my implementation of Transformer.] [Updated on 2018-11-06.
Why it mattersA rigorous, implementation-backed walkthrough of attention and the Transformer architecture that grounds engineers in the mechanics behind modern LLMs, with code links for self-attention, Pointer Networks, and Neural Turing Machines.
articleLilian Weng
Implementing Deep Reinforcement Learning Models with Tensorflow + OpenAI Gym
A hands-on implementation walkthrough of deep reinforcement learning models in TensorFlow against OpenAI Gym.
Why it mattersBridges the gap between deep RL theory and working code, walking through actual TensorFlow + OpenAI Gym implementations of models most tutorials only describe abstractly.
articleLilian Weng
Policy Gradient Algorithms
A survey of policy gradient algorithms in reinforcement learning, from the basic theorem through actor-critic variants including SAC, D4PG, TD3 and SVPG.
Why it mattersA comprehensive, math-first survey of policy gradient methods that walks through the derivations and design tradeoffs of REINFORCE, A2C/A3C, DDPG, TD3, SAC, PPO, IMPALA and more in one place.
articleLilian Weng
A (Long) Peek into Reinforcement Learning
[Updated on 2020-09-03: Updated the algorithm of SARSA and Q-learning so that the difference is more pronounced. [Updated on 2021-09-19.
Why it mattersA rigorous, foundational walkthrough of core RL algorithms (SARSA, Q-learning, policy gradients) that grounds the concepts increasingly relevant to agent training and RLHF-style fine-tuning.
articleLilian Weng
The Multi-Armed Bandit Problem and Its Solutions
The exploration-exploitation dilemma stated as the multi-armed bandit problem.
Why it mattersA clear, code-backed walkthrough of bandit algorithms (epsilon-greedy, UCB, Thompson sampling) that maps the exploration/exploitation tradeoff to real problems like ad selection and A/B testing.
articleLilian Weng
Learning Word Embedding
Human vocabulary comes in free text.
Why it mattersA clear walkthrough of how free-text words become numeric vectors — from one-hot encoding to learned embeddings — grounding the intuition behind the embedding models and vector search that agentic engineers rely on daily.
articleLilian Weng
Anatomize Deep Learning with Information Theory
Naftali Tishby's information bottleneck applied to deep learning, using information theory to describe how a network's representations grow and transform over the course of training.
Why it mattersIt unpacks Tishby's Information Bottleneck framework and the two-phase (fitting then compression) view of DNN training, giving practitioners an information-theoretic lens on generalization that most engineering-focused writeups skip.
articleLilian Weng
From GAN to WGAN
A walk from the original generative adversarial network to Wasserstein GAN.
Why it mattersA rigorous, well-illustrated walkthrough of why vanilla GANs are unstable and how Wasserstein distance fixes the gradient/convergence problems — useful grounding for anyone building or debugging generative models.
articleLilian Weng
An index of the vibe-coding frontier. Corrections welcome.