Model · phi-2

phi-2

Developer: Microsoft
Availability: available · checked 2026-09-24 · source

Raw record: /data/models/phi-2.json

Fields

id

phi-2

identifiers

huggingface

microsoft/phi-2

developer

Microsoft

release_date

2023-12-13partial · source

note:  HF repo creation date; may predate or postdate public release.

weights_status

open

availability

available · checked 2026-09-24 · source

license

MITrecorded · source

architecture

family

decoder_only

note

family read from config 'architectures': ['PhiForCausalLM'].

n_layers

32recorded · source

hidden_size

2560recorded · source

n_heads

32recorded · source

vocab_size

51200recorded · source

context_length

2048recorded · source

n_kv_heads

32recorded · source

positional_encoding

nullnot_recorded

training_data

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.recorded · source

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 ·

Parents

Training data

• trained_on → RefinedWeb declared source Card: training mix includes 'filtered web data from Falcon RefinedWeb and SlimPajama, which was assessed by AOAI GPT-4'.

Children

No edges recorded.

Read in

No station on the reading path has touched this record yet.