repoGitHubTrust 82 · PrimaryPublished 1mo agoLive · 1mo ago

rasbt/reasoning-from-scratch

Implement a reasoning LLM in PyTorch from scratch, step by step

Lineage graph

Paper → model → repo connections mined from source citations (Tier-1 exact match).

Why these links exist

Every edge carries a method, confidence, and the source snippet that justified it — so bad links are debuggable.

PossiblePossibly related (embedding) · 53%I wrote a GGUF inferencer from scratch, AMA →
PossiblePossibly related (embedding) · 52%H64LM: A 249M-parameter Mixture-of-Experts Transformer built from scratch in PyTorch [P] →
PossiblePossibly related (embedding) · 48%Profiling in PyTorch (Part 2): From nn.Linear to a Fused MLP →
PossiblePossibly related (embedding) · 48%ThinkProbe: Beyond Accuracy -- Structural Profiling of Open-Ended LLM Reasoning Traces via Non-Generative Thought Graphs →
PossiblePossibly related (embedding) · 48%COCOLogic-V2: Identifying Logical Inconsistencies via Truly Hard-Negatives →
FuzzySimilar title/name (fuzzy) · 59%Towards Enhancing 3D Spatial Reasoning in Medical Multimodal Large Language Models →
“Fuzzy title match (0.73): “Towards Enhancing 3D Spatial Reasoning in Medical Multimodal” ≈ “rasbt/reasoning-from-scratch””
FuzzySimilar title/name (fuzzy) · 59%Learning Mechanistic Reasoning for Chemical Reactions with Large Language Models →
“Fuzzy title match (0.73): “Learning Mechanistic Reasoning for Chemical Reactions with L” ≈ “rasbt/reasoning-from-scratch””
FuzzySimilar title/name (fuzzy) · 59%SCOPE-RL: Optimizing Reasoning Paths Before and After Success →
“Fuzzy title match (0.73): “SCOPE-RL: Optimizing Reasoning Paths Before and After Succes” ≈ “rasbt/reasoning-from-scratch””

Covers

newsI wrote a GGUF inferencer from scratch, AMA newsH64LM: A 249M-parameter Mixture-of-Experts Transformer built from scratch in PyTorch [P]newsProfiling in PyTorch (Part 2): From nn.Linear to a Fused MLP

Implements

paperThinkProbe: Beyond Accuracy -- Structural Profiling of Open-Ended LLM Reasoning Traces via Non-Generative Thought Graphs paperCOCOLogic-V2: Identifying Logical Inconsistencies via Truly Hard-Negatives paperTowards Enhancing 3D Spatial Reasoning in Medical Multimodal Large Language Models paperLearning Mechanistic Reasoning for Chemical Reactions with Large Language Models paperSCOPE-RL: Optimizing Reasoning Paths Before and After Success paperLearning to Evolve Scenes: Reasoning about Human Activities with Scene Graphs paper4DR360: State Reasoning for Joint 3D Detection and Occupancy Prediction in 4D Radar-Camera Full-Scene Perception paperDiffusion-GR2: Diffusion Generative Reasoning Re-ranker paperReContext: Recursive Evidence Replay as LLM Harness for Long-Context Reasoning paperWILDTRACE: Benchmarking Natural Evidence Trails in Long-Context Reasoning paperMulti-scale Object-Aware Gaze Estimation via Geometric Reasoning paperToken-Sparse Medical Multimodal Reasoning via Dual-Stream Reinforcement Learning paperG-RRM: Guiding Symbolic Solvers with Recurrent Reasoning Models paperBefore Thinking, Learn to Decide: Proactive Routing for Efficient Visual Reasoning paperVisual Access Boundaries in Vision-Language Model Reasoning paperEstimating Uncertainty from Reasoning: A Large-Scale Study of Multi- and Crosslingual MCQA Performance in LLMs paperThe Complexity Ceiling Benchmark: A Multi-Domain Evaluation of Sequential Reasoning Under Depth Scaling paperVerifier-Based Reinforcement Fine-Tuning of Reasoning Models for Thermal Energy Storage Control paperIdeas Have Genomes: Benchmarking Scientific Lineage Reasoning and Lineage-Grounded Idea Generation paperDT-Guard: Intent-Driven Reasoning-Active Training for Reasoning-Free LLM Safety Guardrail paperCLExEval: A Human-in-the-Loop Framework for Qualitative Evaluation of LLM Clinical Reasoning paperAre We Measuring Strategy or Phrasing? The Gap Between Surface- and Approach-Level Diversity in LLM Math Reasoning paperEchoSonar-R: A Multi-View Reasoning-Enabled Model for Disease Classification and Report Generation in Echocardiography paperTask-Specific Multimodal Question Answering Agents via Confidence Calibration and Incremental Reasoning for QANTA 2026 paperDo Vision-Language-Action Models Mean What They Say? On the Role of Faithfulness in Embodied Reasoning paperInvariant Learning Dynamics of Transformers in Inductive Reasoning Tasks paperReaORE: Reasoning-Guided Progressive Open Relation Extraction Empowered by Large Reasoning Models paperDeep Interaction: An Efficient Human-AI Interaction Method for Large Reasoning Models paperShow Me How You Reason and I'll Tell You Who You Are: Reasoning Graphs for Robust LLM Authorship Attribution paperCoTu at EXACT 2026: Neuro-Symbolic Reasoning for Transparent Educational QA paperGold-Guided Programmatic Distillation for Financial Reasoning over Hybrid Tables and Text paperStop Thinking, Start Looking: Efficient Post-Training for Multimodal Document Question Answering via Reasoning-Free Alignment paperLeveraging Instruction Tuning and Merging for Reasoning Model Adaptation paperHierarchical Denoising For Multi-Step Visual Reasoning paperHoloGeo: Mitigating Landmark Bias in Geo-localization via Evidence-Driven Reasoning paperDetecting Answer-Driven Reasoning in LLM-Based Educational Tutors via Truncated Chain-of-Thought Auditing paperLearning to Select, Not Relearn: Hard-Routed Mixtures of Reasoning LoRAs paperDo AI Agents Know When a Task Is Simple? Toward Complexity-Aware Reasoning and Execution paperA rubric-based controlled comparison of frontier language models on expert-authored clinical reasoning tasks

Covers (incoming)

newsProfiling in PyTorch (Part 3): Attention is all you profile

Implements (incoming)

paperThink Through a Bottleneck: Hourglass Reasoning for Rigorous Induction

Related to (incoming)

modelJackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled

Related across the graph

paper4DR360: State Reasoning for Joint 3D Detection and Occupancy Prediction in 4D Radar-Camera Full-Scene Perception paperDiffusion-GR2: Diffusion Generative Reasoning Re-ranker paperSCOPE-RL: Optimizing Reasoning Paths Before and After Success paperReContext: Recursive Evidence Replay as LLM Harness for Long-Context Reasoning newsProfiling in PyTorch (Part 2): From nn.Linear to a Fused MLP paperLearning Mechanistic Reasoning for Chemical Reactions with Large Language Models paperVerifier-Based Reinforcement Fine-Tuning of Reasoning Models for Thermal Energy Storage Control paperWILDTRACE: Benchmarking Natural Evidence Trails in Long-Context Reasoning paperMulti-scale Object-Aware Gaze Estimation via Geometric Reasoning paperToken-Sparse Medical Multimodal Reasoning via Dual-Stream Reinforcement Learning paperCOCOLogic-V2: Identifying Logical Inconsistencies via Truly Hard-Negatives paperG-RRM: Guiding Symbolic Solvers with Recurrent Reasoning Models paperLearning to Evolve Scenes: Reasoning about Human Activities with Scene Graphs paperBefore Thinking, Learn to Decide: Proactive Routing for Efficient Visual Reasoning paperHierarchical Denoising For Multi-Step Visual Reasoning paperVisual Access Boundaries in Vision-Language Model Reasoning newsProfiling in PyTorch (Part 3): Attention is all you profile paperEstimating Uncertainty from Reasoning: A Large-Scale Study of Multi- and Crosslingual MCQA Performance in LLMs paperThe Complexity Ceiling Benchmark: A Multi-Domain Evaluation of Sequential Reasoning Under Depth Scaling paperHoloGeo: Mitigating Landmark Bias in Geo-localization via Evidence-Driven Reasoning paperIdeas Have Genomes: Benchmarking Scientific Lineage Reasoning and Lineage-Grounded Idea Generation paperDT-Guard: Intent-Driven Reasoning-Active Training for Reasoning-Free LLM Safety Guardrail paperAre We Measuring Strategy or Phrasing? The Gap Between Surface- and Approach-Level Diversity in LLM Math Reasoning paperStop Thinking, Start Looking: Efficient Post-Training for Multimodal Document Question Answering via Reasoning-Free Alignment paperCLExEval: A Human-in-the-Loop Framework for Qualitative Evaluation of LLM Clinical Reasoning paperDo AI Agents Know When a Task Is Simple? Toward Complexity-Aware Reasoning and Execution paperDo Vision-Language-Action Models Mean What They Say? On the Role of Faithfulness in Embodied Reasoning paperEchoSonar-R: A Multi-View Reasoning-Enabled Model for Disease Classification and Report Generation in Echocardiography paperDetecting Answer-Driven Reasoning in LLM-Based Educational Tutors via Truncated Chain-of-Thought Auditing paperLearning to Select, Not Relearn: Hard-Routed Mixtures of Reasoning LoRAs paperDeep Interaction: An Efficient Human-AI Interaction Method for Large Reasoning Models paperInvariant Learning Dynamics of Transformers in Inductive Reasoning Tasks modelJackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled paperReaORE: Reasoning-Guided Progressive Open Relation Extraction Empowered by Large Reasoning Models paperTask-Specific Multimodal Question Answering Agents via Confidence Calibration and Incremental Reasoning for QANTA 2026 paperA rubric-based controlled comparison of frontier language models on expert-authored clinical reasoning tasks paperThink Through a Bottleneck: Hourglass Reasoning for Rigorous Induction newsH64LM: A 249M-parameter Mixture-of-Experts Transformer built from scratch in PyTorch [P]paperThinkProbe: Beyond Accuracy -- Structural Profiling of Open-Ended LLM Reasoning Traces via Non-Generative Thought Graphs paperLeveraging Instruction Tuning and Merging for Reasoning Model Adaptation paperTowards Enhancing 3D Spatial Reasoning in Medical Multimodal Large Language Models paperShow Me How You Reason and I'll Tell You Who You Are: Reasoning Graphs for Robust LLM Authorship Attribution paperGold-Guided Programmatic Distillation for Financial Reasoning over Hybrid Tables and Text paperCoTu at EXACT 2026: Neuro-Symbolic Reasoning for Transparent Educational QA newsI wrote a GGUF inferencer from scratch, AMA

Topics

ai artificial-intelligence chain-of-thought deep-learning distillation grpo inference-time-scaling large-language-models llm llms