Study the field in the order its ideas emerged. Each paper includes guided notes, focused learning modules, and questions that test real understanding.
70Guided papers
180Learning modules
2017 — 2026Research span
RESEARCH PATH
Read in sequence.
DATE
NO. / YEARPAPERRESEARCH AREAACCESS / PROGRESS
012017V100Research paper · 1 learning moduleHardware022017RLHFDeep reinforcement learning from human preferences · 1 learning moduleAlignment032017TransformerAttention Is All You Need · 4 learning modulesArchitecture042017PPOProximal Policy Optimization Algorithms · 1 learning moduleAlignment052017QATQuantization and Training of Neural Networks for Efficient Integer-Arithmetic-Only Inference · 1 learning moduleQuantization062018Top-kHierarchical Neural Story Generation · 1 learning moduleInference072018GPT-1Improving Language Understanding by Generative Pre-Training · 7 learning modulesLanguage ModelingFine-Tuning082018Tesla T4NVIDIA · 1 learning moduleHardware092018BERTBERT: Pre-training of Deep Bidirectional Transformers for Language Understanding · 7 learning modulesLanguage ModelingFine-Tuning102018GPipeGPipe: Efficient Training of Giant Neural Networks using Pipeline Parallelism · 1 learning moduleSystems112018Hugging FaceHugging Face · 1 learning moduleTooling122019GPT-2Language Models are Unsupervised Multitask Learners · 5 learning modulesLanguage Modeling132019Top-pThe Curious Case of Neural Text Degeneration · 1 learning moduleInference142019Tensor ParallelMegatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism · 1 learning moduleSystems152019Data ParallelZeRO: Memory Optimizations Toward Training Trillion Parameter Models · 1 learning moduleSystems162019T5Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer · 7 learning modulesLanguage Modeling172019MQAFast Transformer Decoding: One Write-Head is All You Need · 1 learning moduleArchitectureInference182020Scaling LawScaling Laws for Neural Language Models · 1 learning moduleScalingLanguage Modeling192020A100Research paper · 1 learning moduleHardware202020GPT-3Language Models are Few-Shot Learners · 5 learning modulesScalingLanguage Modeling212020GShardGShard: Scaling Giant Models with Conditional Computation and Automatic Sharding · 1 learning moduleSystemsArchitecture222020ViTAn Image is Worth 16x16 Words: Transformers for Image Recognition at Scale · 3 learning modulesArchitectureVision232021Switch TransformerSwitch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsity · 1 learning moduleArchitectureScaling242021ZeRO-OffloadZeRO-Offload: Democratizing Billion-Scale Model Training · 1 learning moduleSystems252021ViLTViLT: Vision-and-Language Transformer Without Convolution or Region Supervision · 3 learning modulesVisionMultimodal262021DALL·E 1Zero-Shot Text-to-Image Generation · 6 learning modulesMultimodalImage Generation272021CLIPLearning Transferable Visual Models From Natural Language Supervision · 3 learning modulesVisionMultimodal282021Megatron-LM: 3D ParallelismEfficient Large-Scale Language Model Training on GPU Clusters Using Megatron-LM · 1 learning moduleSystems292021ZeRO-InfinityZeRO-Infinity: Breaking the GPU Memory Wall for Extreme Scale Deep Learning · 1 learning moduleSystems302021LoRALoRA: Low-Rank Adaptation of Large Language Models · 1 learning moduleFine-Tuning312021CodeXEvaluating Large Language Models Trained on Code · 4 learning modulesLanguage ModelingCode322021GLaMGLaM: Efficient Scaling of Language Models with Mixture-of-Experts · 1 learning moduleArchitectureScaling332021Stable DiffusionHigh-Resolution Image Synthesis with Latent Diffusion Models · 2 learning modulesMultimodalImage Generation342022AlphaCodeCompetition-Level Code Generation with AlphaCode · 5 learning modulesReasoningCode352022COTChain-of-Thought Prompting Elicits Reasoning in Large Language Models · 1 learning moduleInferenceReasoning362022InstructGPTTraining language models to follow instructions with human feedback · 5 learning modulesAlignmentFine-Tuning372022H100Research paper · 1 learning moduleHardware382022ChinchillaTraining Compute-Optimal Large Language Models · 1 learning moduleScalingLanguage Modeling392022DALL·E 2Hierarchical Text-Conditional Image Generation with CLIP Latents · 3 learning modulesMultimodalImage Generation402022FlashAttentionFlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness · 1 learning moduleSystemsInference412022LLM.int8()LLM.int8(): 8-bit Matrix Multiplication for Transformers at Scale · 1 learning moduleQuantizationInference422022GPTQGPTQ: Accurate Post-Training Quantization for Generative Pre-trained Transformers · 1 learning moduleQuantizationInference432022WhisperRobust Speech Recognition via Large-Scale Weak Supervision · 2 learning modulesSpeech442023LLaMA-1LLaMA: Open and Efficient Foundation Language Models · 5 learning modulesScalingLanguage Modeling452023LLaVAVisual Instruction Tuning · 2 learning modulesVisionMultimodal462023DPODirect Preference Optimization: Your Language Model is Secretly a Reward Model · 1 learning moduleAlignmentFine-Tuning472023QLoRAQLoRA: Efficient Finetuning of Quantized LLMs · 1 learning moduleFine-TuningQuantization482023AWQAWQ: Activation-aware Weight Quantization for LLM Compression and Acceleration · 1 learning moduleQuantizationInference492023FlashAttention-2FlashAttention-2: Faster Attention with Better Parallelism and Work Partitioning · 1 learning moduleSystemsInference502023LLaMA-2Llama 2: Open Foundation and Fine-Tuned Chat Models · 7 learning modulesLanguage ModelingAlignment512023Qwen-VLQwen-VL: A Versatile Vision-Language Model for Understanding, Localization, Text Reading, and Beyond · 3 learning modulesVisionMultimodal522023Qwen-1Qwen Technical Report · 7 learning modulesLanguage Modeling532023Mistral 7BMistral 7B · 5 learning modulesArchitectureInference542023H200NVIDIA H200 Tensor Core GPU Datasheet · 1 learning moduleHardware552023LVMSequential Modeling Enables Scalable Learning for Large Vision Models · 3 learning modulesArchitectureVision562024KTOKTO: Model Alignment as Prospect Theoretic Optimization · 1 learning moduleAlignmentFine-Tuning572024Mixtral 8x7BMixtral of Experts · 2 learning modulesArchitectureScaling582024ORPOORPO: Monolithic Preference Optimization without Reference Model · 1 learning moduleAlignmentFine-Tuning592024Gemma 1Gemma: Open Models Based on Gemini Research and Technology · 7 learning modulesLanguage Modeling602024DeepSeek-V2DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model · 2 learning modulesArchitectureInference612024SimPOSimPO: Simple Preference Optimization with a Reference-Free Reward · 1 learning moduleAlignmentFine-Tuning622024B200Research paper · 1 learning moduleHardware632024ChatGLMChatGLM: A Family of Large Language Models from GLM-130B to GLM-4 All Tools · 3 learning modulesLanguage Modeling642024Llama 3The Llama 3 Herd of Models · 6 learning modulesScalingLanguage Modeling652024Gemma 2Gemma 2: Improving Open Language Models at a Practical Size · 6 learning modulesArchitectureLanguage Modeling662024DeepSeek-V3DeepSeek-V3 Technical Report · 4 learning modulesSystemsArchitecture672025DeepSeek-R1DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning · 3 learning modulesLanguage ModelingReasoning682025B300NVIDIA GTC 2025 Keynote & Blackwell Ultra Architecture Brief · 1 learning moduleHardware692025Gemma 3Gemma 3 Technical Report · 6 learning modulesLanguage ModelingMultimodal702026Rubin GPUNVIDIA Vera Rubin NVL72 Product Page · 1 learning moduleHardware
FULL RESEARCH ACCESS
One library. Every guided paper.
Unlock the complete reading path, every learning module, personal notes, quizzes, and future paper updates.
