--- dataset_info: features: - name: messages list: - name: content dtype: string - name: role dtype: string - name: source dtype: string splits: - name: train num_bytes: 42141633724 num_examples: 6063768 download_size: 14226339862 dataset_size: 42141633724 configs: - config_name: default data_files: - split: train path: data/train-* --- # Instruct Data From Various Sources | Source | Num Samples | Language | Remarks | |-----------------------------------------------------|-------------|-----------|----------------------------------------------| | BAAI/IndustryInstruction | 897.2k | en | Filtered Chinese | | PrimeIntellect/INTELLECT-MATH-SFT-Data | 733k | en | null | | BhabhaAI/openhermes-2.5-hindi | 620.2k | hi | null | | anirudhlakhotia/Baarat-Hindi-Instruct-v1 | 573k | hi | null | | KodCode/KodCode-V1 | 444k | en | null | | GeneralReasoning/GeneralThought-430K | 431k | en | null | | smol-magpie-ultra-short | 270.8k | en | null | | dnyanesh/HindiMathQuest | 250.1k | hi | null | | mlabonne/opc-sft-stage2-chat | 224.5k | en | null | | manishiitg/hieunguyenminh-roleplay | 159.2k | hi | null | | manishiitg/HuggingFaceH4-ultrachat_200k | 147.4k | hi | null | | damerajee/camel-phy-chem-bio-sharegpt | 120k | hi-en mix | null | | OpenThoughts-114k-math-default | 114k | en | null | | sarvamai/samvaad-hi-v1 | 101.4k | hi | null | | Ornaments/cot-hindi-hinglish | 60.1k | hi | null | | mlabonne/lmsys-arena-human-sft-55k | 57.4k | en | null | | open-r1/SYNTHETIC-1-SFT-Data-Code_decontaminated | 49.7k | en | null | | open-r1/codeforces-cots | 47.8k | en | null | | nothingiisreal/Kalomaze-Opus-Instruct-25k-filtered | 48.7k | en | Made BiBo-Specific | | self-oss-instruct | 48.1k | en | null | | openhermes-50k | 47.5k | en | null | | smol-contraints | 34.4k | en | null | | allenai/tulu-3-sft-personas-instruction-following | 30k | en | null | | fhai50032/Hindi-Instruct-HQ | 27.7k | hi | Filtered Small and Hallucinated Response | | smollm-rewrite-30k | 26.7k | en | null | | smol-summarize-20k | 19.3k | en | null | | data/PoT/mathqa.json | 16.3k | en | null | | aaditya/databricks-dolly-15k-Hinglish-Codemix | 15k | en/hi | null | | one-thing/chatbot_arena_conversations_hinglish | 11.3k | en/hi | null | | BhabhaAI/Hi-Instruct-v0 | 10k | hi | null | | HuggingFaceH4/deita-10k-v0-sft | 10k | en | null | | data/PoT/gsm_gpt4.json | 9.7k | en | null | | fhai50032/Hinglish-Preference-Humanized-DPO | 9.4k | en/hi | english question - chosen hinglish answer | | fhai50032/Hinglish-Preference-Humanized-DPO | 9.4k | en/hi | hinglish question - chosen hinglish answer | | data/PoT/numglue.json | 8.9k | en | null | | fhai50032/Hinglish-Preference-Humanized-DPO | 6.9k | en/hi | hinglish question - rejected hinglish answer | | fhai50032/Hinglish-Preference-Humanized-DPO | 6.9k | en/hi | english question - rejected hinglish answer | | data/PoT/aqua_rat_filtered.json | 6.4k | en | null | | smol-summarize-5k | 4.7k | en | null | | longalign | 3.6k | en | null | | macadeliccc/opus_samantha | 3.2k | en | Specified To BiBo | | explore-instruct-rewrite | 3k | en | null | | HuggingFaceTB/everyday-conversations-llama3.1-2k | 2.4k | en | null | | everyday-conversations | 2.3k | en | null | | digitalpipelines/samantha-1.1-uncensored | 2k | en | null | | OdiaGenAI/roleplay_hindi | 1k | hi | null | | smangrul/hinglish_self_instruct_v0 | 1k | en/hi | null | | Abhaykoul/Ancient-Indian-Wisdom | 0.6k | en | null | | GAIR/o1-journey | 0.3k | en | null | | fhai50032/Aryabhatt-Hinglish | 0.2k | en/hi | CoT | | fhai50032/Aryabhatt-Hinglish | 0.2k | en | CoT | # Language Metrics From the table above (56 rows total), we have three language categories: - **Pure English ("en")**: 39 rows Total samples ≈ 3,618.8k (3,618,800) - **Pure Hindi ("hi")**: 9 rows Total samples ≈ 2,051.5k (2,051,500) - **Mixed (e.g., "en/hi", "hi-en mix")**: 9 rows Total samples ≈ 180.1k (180,100) **Overall total samples** ≈ 5,850.4k (5,850,400) ### Percentages: - **Pure English**: ~ 61.85% - **Pure Hindi**: ~ 35.07% - **Mixed**: ~ 3.08%