{
  "id": "phi-2",
  "name": "phi-2",
  "identifiers": {
    "huggingface": [
      "microsoft/phi-2"
    ]
  },
  "developer": "Microsoft",
  "release_date": {
    "value": "2023-12-13",
    "source": "https://huggingface.co/api/models/microsoft/phi-2",
    "status": "partial",
    "note": "HF repo creation date; may predate or postdate public release."
  },
  "weights_status": "open",
  "availability": {
    "value": "available",
    "checked": "2026-09-24",
    "source": "https://huggingface.co/microsoft/phi-2"
  },
  "license": {
    "value": "MIT",
    "source": "https://huggingface.co/microsoft/phi-2",
    "status": "recorded"
  },
  "architecture": {
    "family": "decoder_only",
    "note": "family read from config 'architectures': ['PhiForCausalLM'].",
    "n_layers": {
      "value": 32,
      "source": "https://huggingface.co/microsoft/phi-2/raw/810d367871c1d460086d9f82db8696f2e0a0fcd0/config.json",
      "status": "recorded"
    },
    "hidden_size": {
      "value": 2560,
      "source": "https://huggingface.co/microsoft/phi-2/raw/810d367871c1d460086d9f82db8696f2e0a0fcd0/config.json",
      "status": "recorded"
    },
    "n_heads": {
      "value": 32,
      "source": "https://huggingface.co/microsoft/phi-2/raw/810d367871c1d460086d9f82db8696f2e0a0fcd0/config.json",
      "status": "recorded"
    },
    "vocab_size": {
      "value": 51200,
      "source": "https://huggingface.co/microsoft/phi-2/raw/810d367871c1d460086d9f82db8696f2e0a0fcd0/config.json",
      "status": "recorded"
    },
    "context_length": {
      "value": 2048,
      "source": "https://huggingface.co/microsoft/phi-2/raw/810d367871c1d460086d9f82db8696f2e0a0fcd0/config.json",
      "status": "recorded"
    },
    "n_kv_heads": {
      "value": 32,
      "source": "https://huggingface.co/microsoft/phi-2/raw/810d367871c1d460086d9f82db8696f2e0a0fcd0/config.json",
      "status": "recorded"
    },
    "positional_encoding": {
      "value": null,
      "status": "not_recorded"
    }
  },
  "training_data": {
    "value": "250B tokens for this training phase (1.4T total across the Phi series): NLP synthetic data from Azure OpenAI GPT-3.5, plus filtered RefinedWeb and SlimPajama web data assessed by GPT-4.",
    "source": "https://huggingface.co/microsoft/phi-2",
    "status": "recorded"
  },
  "techniques": [],
  "primary_sources": [
    "https://huggingface.co/microsoft/phi-2",
    "https://huggingface.co/microsoft/phi-2/raw/810d367871c1d460086d9f82db8696f2e0a0fcd0/config.json"
  ],
  "record_history": [
    {
      "date": "2026-09-24",
      "change": "ingested as candidate from HF (microsoft/phi-2@810d367871c1d460086d9f82db8696f2e0a0fcd0)",
      "by": "ingest_hf.py"
    },
    {
      "date": "2026-09-24",
      "change": "preparer: developer, license, trained_on refinedweb (partial) added, training_data filled",
      "by": "claude (preparer, Sonnet 5)"
    },
    {
      "date": "2026-09-24",
      "change": "preparer cleanup: removed 2 stale flag(s)",
      "by": "claude (preparer, Sonnet 5)"
    },
    {
      "date": "2026-09-24",
      "change": "reviewed and promoted from staging (1 edge(s) accepted)",
      "by": "Wilson Pruitt"
    }
  ]
}