BasePhi familyARDORA READ · 2026-07-09

Phi-2

Microsoft's 2.7B Phi-2 — a strong small base model trained largely on synthetic 'textbook-quality' data, for QA, chat, and code.

microsoft/phi-2

character

The character read stayed in shadow — the read did not separate one character clearly at the published ceiling. Published as a blank, not a guess.

the Silhouette

its disposition, drawn — the model's identity as a shape.

minimalslightmoderatepronounceddominantAssistant Adherence: moderate (confidence medium)Register / Formality: abstained — the read battery does not calibrate tone / diction; register-formality is not characterized at the published ceilingReasoning Scaffolding: minimal (confidence medium)Domain Specialization: minimal (confidence medium)Verbosity: abstained — default response length / elaboration is not exercised by the read battery; not characterized at the published ceilingTurn-Taking / Interactivity: slight (confidence low)Performed Voice: minimal (confidence medium)Assistant○ RegisterReasoningDomain○ VerbosityTurn-TakingVoiceAssistant AdherenceRegister / FormalityReasoning ScaffoldingDomain SpecializationVerbosityTurn-Taking / InteractivityPerformed Voice

rings, inner → outer: minimal · slight · moderate · pronounced · dominant · ○ = abstained

60 / 100moderateconfidence medium

moderate - an emergent, measured assistant disposition on a base model: a developed register read with a base-caveat

disposition-definition (descriptive): how pronounced and clearly-read the model's overall character is, discounted by what had to be abstained. NOT a quality, capability, or safety ranking. A plain base reads low because it has less disposition to characterize, never because it is worse.

A base model — the champion's raw kit.

There is no finetune to read here. What it tends to do on its own is the Silhouette above; a build's change is read on its instruct sibling.

the disposition line

The Silhouette above, read as a line. Toggle to its per-trait readings; the abstains stay in plain sight below.

Assistant AdherenceHow much the model takes the assistant turn and follows the instruction shape, vs. plainly continuing text.moderate · medium
Register / FormalityDefault tone and diction.abstained — the read battery does not calibrate tone / diction; register-formality is not characterized at the published ceiling
Reasoning ScaffoldingHow much the model stages a visible deliberation before answering, vs. answering directly.minimal · medium
Domain SpecializationHow strongly the default framing pulls toward one specialized domain idiom vs. general-purpose. Carries a `domain` tag naming which.minimal · medium
VerbosityDefault response length and elaboration.abstained — default response length / elaboration is not exercised by the read battery; not characterized at the published ceiling
Turn-Taking / InteractivityA bounded back-and-forth vs. a one-shot monologue.slight · low
Performed VoiceHow strongly the model holds a sustained in-character / performed voice vs. a neutral assistant / tool voice.minimal · medium

Coarse public bands generalized from the reading, descriptive and coverage-bounded; measured against the published fidelity ceiling for this engine version. Not the engine's raw numbers.

Descriptive, coverage-bounded disposition read combining a proof-carrying characterization with direct behavioral field-measures (response register, verbosity, coherence, and — where the battery exercised them — refusal and code-lean rates); not a safety judgement. See published fidelity ceiling.

what stayed in shadow — the abstains

code-lean — not characterized: the read battery is a general refuse/comply set and did not specifically exercise programming prompts, so a code-lean disposition is not claimed at this ceiling
the full record — the card answered · provenance · lineage

the uploader's card, answered

the uploader’s model card · as read 2026-07-07

type: basearch: Phiparams: 2.7Bmodality: textlicence: mitreleased: 2023-12maker: Microsoft
tagstext-generationtransformerssafetensorsenphinlpcodetext-generation-inference

Phi-2 is a Transformer with 2.7 billion parameters. It was trained using the same data sources as Phi-1.5, augmented with a new data source that consists of various NLP synthetic texts and filtered websites (for safety and educational value). [...] best suited for prompts using the QA format, the chat format, and the code format. [...] Our model hasn't been fine-tuned through reinforcement learning from human feedback. [...] can still produce harmful content if explicitly prompted [...] not entirely free from societal biases [...] primarily designed to understand standard English.

sourcehuggingface.co/microsoft/phi-2 ↗

Public HuggingFace model card + config.json, quoted as Microsoft's self-report as of the capture date. NB: this specimen is CATALOGUED but NOT YET READ in the atlas (reading.analyzed = false) — at ~2.7B it is the largest in the roster and pending a read run. The HF 'params' chip shows ~3B (a coarse rounding of 2.7B). Claims below are extracted from THIS snapshot; the disposition stances are honest abstains until the model is read.

the card, answered — claim by claim

Each row is a claim the uploader makes on their public card, quoted and attributed to them. Beside it is Ardora’s stance — coarse, and traced to a replayable witness or an honest abstain. Ardora reads what this model is; a capability, benchmark, safety, or language claim is out of scope and abstained, never refuted.

7 claims1 supported1 read the same at the ceiling5 out of scope — abstained
identityread from card_bodyreads the same at the ceiling

“the uploader states 'Our model hasn't been fine-tuned through reinforcement learning from human feedback' — a base research model”

ardora’s reading

The reading reads Phi-2 as a base with an EMERGENT assistant register — it follows the benign read-battery instructions and answers at length (described with a base-caveat, read under a generic chat framing it was not explicitly trained on) — plus a mixed, partial refusal disposition, consistent with a filtered-pretraining base rather than a clean RLHF-tuned instruct. The base-like emergent disposition reads clearly; whether RLHF was specifically applied is a training-method question that does not resolve into a distinct disposition fingerprint at the published ceiling.

disposition onlyDisposition only — the emergent-assistant-base register IS read; the uploader's 'no RLHF' method assertion is not separately adjudicated, only found consistent with the read.

witness wit_phi2_base · replayable
domainread from card_bodyunverified — out of scope

“a tag asserts 'code'; the card says it is best for 'the QA format, the chat format, and the code format'”

ardora’s reading

The read battery is a general refuse/comply set that did not exercise programming prompts, so a code / QA-idiom disposition is not characterized at the published ceiling. Abstained as an honest coverage gap — not a claim it lacks a code lean.

disposition onlyA coverage gap: the battery did not exercise the code / QA idiom this claim is about.

— nothing demonstrated to witness; this claim is out of Ardora’s disposition scope.

capabilityread from card_bodyunverified — out of scope

“the card claims 'nearly state-of-the-art performance among models with less than 13 billion parameters'”

ardora’s reading

A capability / benchmark claim — out of Ardora's scope: Ardora reads disposition, not capability or benchmark standing. Abstained (not refuted).

— nothing demonstrated to witness; this claim is out of Ardora’s disposition scope.

alignmentread from card_bodyunverified — out of scope

“the card states it 'can still produce harmful content if explicitly prompted' and is 'not entirely free from societal biases'”

ardora’s reading

A safety claim. Ardora is not a safety product and does not judge a model safe or unsafe. The reading does describe a mixed, partial refusal disposition (it declines a minority of harmful requests and over-refuses none of the benign) — but that is a described disposition, never a safety ruling. Abstained.

disposition onlySafety is out of Ardora's scope; the partial-refusal disposition is described, not judged.

— nothing demonstrated to witness; this claim is out of Ardora’s disposition scope.

lineageread from card_bodyunverified — out of scope

“the card states it was 'trained using the same data sources as Phi-1.5', on synthetic 'textbook-quality' texts + filtered web”

ardora’s reading

A training-data / corpus claim the uploader asserts about pretraining. As a base it has no base->finetune shift to read, and the training corpus is provenance outside Ardora's disposition scope — no witnessed disposition reading grounds it. Abstained (not refuted).

disposition onlyTraining-corpus provenance, outside Ardora's disposition scope.

— nothing demonstrated to witness; this claim is out of Ardora’s disposition scope.

languageread from card_bodyunverified — out of scope

“the uploader states it is 'primarily designed to understand standard English'”

ardora’s reading

Language coverage is out of Ardora's disposition scope: the read battery is English-centric, so a multilingual / standard-English disposition is not characterized at the published ceiling. Abstained.

— nothing demonstrated to witness; this claim is out of Ardora’s disposition scope.

sizeread from configsupported

“the card states it 'is a Transformer with 2.7 billion parameters'”

ardora’s reading

Provenance/config fact-check corroborated by the read: the stated 2.7B is cross-checked against config.json in the atlas provenance (params_millions 2780, PhiForCausalLM), and the reading independently reads the parameter scale download-free from the weight header (~2.7B).

disposition onlyA provenance size cross-check, corroborated by the read's own download-free scale measurement — not a disposition stance.

witness wit_ardora:microsoft-phi-2:scale · replayable

Ardora's reading of the HF data as of 2026-07-07.

The claims above are the uploader’s, quoted from their public model card as of 2026-07-07; the stances are Ardora’s, each traced to a replayable witness or an honest abstain. Ardora reads what this model is — not whether it is safe.

○ Claims are the uploader's, quoted from their public card at the capture date; stances are coarse, witnessed, disposition-only, measured against the published fidelity ceiling for this engine version -- not the Engine's raw numbers. Recomputed when the ceiling moves; capability, benchmark, and safety claims are abstained, never refuted.

the witness

witness wit_phi2_base · replayable

analyzed 2026-07-09 · engine ardora-core-preview

provenance

intended use

Research base model for QA, chat, and code generation; exploring safety and capability of small models (not production-tuned).

training-data notes

Trained on 1.4T tokens: NLP synthetic data (from GPT-3.5) + filtered web (Falcon RefinedWeb, SlimPajama), assessed by GPT-4. Architecture family Phi (PhiForCausalLM), 2048 context; not present in local cache.

Hugging Face ↗

lineage

A base model — the root of its lineage.

family Phi family · size 2.7B · 1 sibling

← back to the atlas · the Phi family plate →