incomplete, I'm still working on this

Tags

This is a map of 28 tags over 16 entries (papers, posts, slides, and posters). Larger nodes mark recurring themes; lines show papers, posts, slides, and posters sharing those themes.

Tags28
Entries (notes) 16
Links64
LargestLLMs

Tag Entry Atlas

Tag and entry networkThe largest tag nodes sit near the center. Outer nodes are entries. Lines connect entries to their tags.Paper: Don’t Pass@π‘˜: A Bayesian Framework for Large Language Model EvaluationPaper: Quantize What Counts: More For Keys, Less For Values β˜οΈπŸ”‘πŸ‘‡πŸ”’Paper: 70% Size, 100% Accuracy: Lossless LLM Compression for Efficient GPU Inference via Dynamic-Length FloatPost: Entropy of bfloat16 During Training: How Optimizers Shape Weight DistributionsPost: Entropy of bfloat16: 8 Bits Are Doing 2.6 Bits of WorkPost: Simulating LLM Evaluation Datasets Using Psychometric ModelsPost: Simulating LLM Answers to Evaluation DatasetsSlide: Python EnvironmentsSlide: LLM Research DirectionsSlide: Virtual Agentic Lab!Slide: Tool Use (Function Calling) & RAGSlide: State of Large Language ModelsSlide: Test-Time Scaling Under BudgetSlide: Psychometric Modeling of LLM Evaluation DatasetsPoster: Ranking Reasoning LLMs under Test-Time ScalingPoster: Don't Pass@k: A Bayesian Framework for Large Language Model EvaluationLLMs: 12 entriesLLMsBayesian: 5 entriesBayesianCompression: 4 entriesCompressionReasoning: 4 entriesReasoningTest-Time Scaling: 4 entriesTest-Time ScalingEfficiency: 3 entriesEfficiencyInference: 3 entriesInferenceSimulation: 3 entriesSimulationBenchmarking: 2 entriesBenchmarkingInformation Theory: 2 entriesInformation TheoryQuantization: 2 entriesQuantizationScorio: 2 entriesScorioStatistics: 2 entriesStatisticsTools: 2 entriesToolsAgents: 1 entriesAgentsAI: 1 entriesAIDatasets: 1 entriesDatasetsEvaluation: 1 entriesGPU: 1 entriesKV Cache: 1 entriesLLM Evaluation: 1 entriesLLM Ranking: 1 entriesLossless: 1 entriesOptimizers: 1 entriesRAG: 1 entriesReasoning Models: 1 entriesTraining: 1 entriesTutorial: 1 entries
Papers Posts Slides Posters

Signal

The strongest tags and pairings show where the archive has accumulated repeated work.

LLMs 3 papers / 3 posts / 6 slides
12 Bayesian 1 papers / 2 slides / 2 posters
5 Compression 1 papers / 2 posts / 1 slides
4 Reasoning 1 papers / 1 posts / 2 slides
4 Test-Time Scaling 1 papers / 1 slides / 2 posters
4 Efficiency 2 papers / 1 posts
3 Inference 2 papers / 1 posts
3 Simulation 1 papers / 2 posts
3 Benchmarking 2 posters
2 Information Theory 2 posts
2
Bayesian + Test-Time Scaling shared entries 4 LLMs + Reasoning shared entries 4 LLMs + Bayesian shared entries 3 LLMs + Compression shared entries 3 LLMs + Efficiency shared entries 3 LLMs + Inference shared entries 3 LLMs + Simulation shared entries 3 Bayesian + Benchmarking shared entries 2

All Tags

Agents 1 AI 1 Bayesian 5 Benchmarking 2 Compression 4 Datasets 1 Efficiency 3 Evaluation 1 GPU 1 Inference 3 Information Theory 2 KV Cache 1 LLM Evaluation 1 LLM Ranking 1 LLMs 12 Lossless 1 Optimizers 1 Quantization 2 RAG 1 Reasoning 4 Reasoning Models 1 Scorio 2 Simulation 3 Statistics 2 Test-Time Scaling 4 Tools 2 Training 1 Tutorial 1