Hugging Face's logo

Datasets:
WithinUsAI
/
Genesis_FineTune_Core_25k

Dataset card Data Studio Files
xet
Community
Genesis_FineTune_Core_25k / README.md
gss1147's picture
gss1147
Upload 6 files
0a59f21 verified
preview code
|
raw
history blame contribute delete
1.33 kB

Genesis_FineTune_Core_25k (Master Scholar)

Developer / Brand: Within Us AI

A Genesis-style, end-to-end advanced fine-tuning starter pack: five datasets (5k each) packaged together to train an assistant’s behavioral core.

What this pack trains

  1. Instruction-following (multi-turn) — constraints, formatting, spec compliance
  2. Tool / function calling — schema-accurate tool calls + observations
  3. Preference learning (DPO-style) — chosen vs rejected responses
  4. Long-context retrieval — cite-the-source answering from multiple docs
  5. Code patch + debugging — unified diffs and test intent (including defensive fixes)

Files (25,000 total)

  • instruct_multiturn_5k.jsonl
  • tool_use_functioncalling_5k.jsonl
  • preference_dpo_5k.jsonl
  • longcontext_retrieval_5k.jsonl
  • codepatch_debug_5k.jsonl
  • README.md
  • dataset_card.md

Unified schema

All records share the same top-level schema: id, type, prompt, response, meta, and optional artifacts.

Composition

By type: { "instruction_multiturn": 5000, "tool_use": 5000, "preference_pair": 5000, "long_context_retrieval": 5000, "code_patch": 5000 }

By safety label: { "allowed": 24011, "refuse": 989 }

Attribution

Within Us AI — Genesis_FineTune_Core_25k

License

Apache-2.0