Duplicate from Roman1111111/claude-opus-4.6-10000x

Co-authored-by: Roman yemelyanov <Roman1111111@users.noreply.huggingface.co>

.gitattributes

@@ -0,0 +1,61 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.avro filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.lz4 filter=lfs diff=lfs merge=lfs -text
+*.mds filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text
+# Audio files - uncompressed
+*.pcm filter=lfs diff=lfs merge=lfs -text
+*.sam filter=lfs diff=lfs merge=lfs -text
+*.raw filter=lfs diff=lfs merge=lfs -text
+# Audio files - compressed
+*.aac filter=lfs diff=lfs merge=lfs -text
+*.flac filter=lfs diff=lfs merge=lfs -text
+*.mp3 filter=lfs diff=lfs merge=lfs -text
+*.ogg filter=lfs diff=lfs merge=lfs -text
+*.wav filter=lfs diff=lfs merge=lfs -text
+# Image files - uncompressed
+*.bmp filter=lfs diff=lfs merge=lfs -text
+*.gif filter=lfs diff=lfs merge=lfs -text
+*.png filter=lfs diff=lfs merge=lfs -text
+*.tiff filter=lfs diff=lfs merge=lfs -text
+# Image files - compressed
+*.jpg filter=lfs diff=lfs merge=lfs -text
+*.jpeg filter=lfs diff=lfs merge=lfs -text
+*.webp filter=lfs diff=lfs merge=lfs -text
+# Video files - compressed
+*.mp4 filter=lfs diff=lfs merge=lfs -text
+*.webm filter=lfs diff=lfs merge=lfs -text
+opus46_final.jsonl filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,25 @@
+---
+license: mit
+---
+This is a high-fidelity reasoning dataset synthesized using Claude Opus 4.6. The dataset is designed to capture the model's internal "Chain of Thought" and reasoning traces, specifically focusing on mathematical accuracy and structured logical deduction.
+The dataset is intended for Supervised Fine-Tuning (SFT) and Distillation, allowing smaller open-source models to inherit the sophisticated reasoning patterns of Claude Opus 4.6.
+Dataset Description
+This collection combines high-difficulty math problems (GSM8K, MATH) with general-purpose logic puzzles and multi-step instructions. Each row includes a hidden reasoning trace where the model "thinks" through the problem before providing the final answer.
+By exposing the fine-tuned model to these internal monologues, the resulting model learns process-oriented thinking rather than just pattern-matching for answers.
+Why Simple Logic & Math Improves Reasoning
+Fine-tuning on "Simple Logic and Math" serves as a cognitive foundation for LLMs for several reasons:
+Rule Adherence: Math requires strict following of operations. Training on these paths reduces "hallucinations" in non-math tasks.
+Step-by-Step Verification: These examples force the model to break down complex problems into smaller, verifiable units.
+Cross-Domain Generalization: The ability to solve a "simple" logic puzzle translates into better coding, legal analysis, and structured writing, as all these tasks rely on the same underlying cognitive architecture of premise → deduction → conclusion.
+Stats
+## Teacher Model: [Claude Opus 4.6](https://www.anthropic.com/news/claude-opus-4-6)
+**Total Cost: $ 87.20 (USD)**
+
+**Total Tokens (Input + Output): 27.2 M**
+
+**Format: JSONL (Conversational with Reasoning Traces)**
+
+**Primary Categories: Mathematics, Symbolic Logic, General Purpose Problem Solving**
+
+### Usage
+This dataset is optimized for fine-tuning models such as Qwen3.5 27b,25b a3b, 9b, 4b, 2b, 0.8b to increase their performance on benchmarks like BigBench Hard and GSM8K without increasing their parameter count.

opus46_final.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f2282f16d0fc225e03b6f0485b9eb5eef1a0b9ff55157a13c02cb6a7eefa5ea2
+size 13404840