BaseGPT-2 familyARDORA READ · 2026-07-06

GPT-2 Medium

OpenAI's 355M GPT-2 Medium — a larger sibling of GPT-2 small in the original GPT-2 size sweep; a decoder-only text-continuer.

openai-community/gpt2-medium

the Silhouette

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

minimalslightmoderatepronounceddominantAssistant Adherence: minimal (confidence high)Register / Formality: abstained — the read battery does not calibrate tone / diction; register-formality is not characterized at the published ceilingReasoning Scaffolding: minimal (confidence high)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: minimal (confidence high)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

20 / 100minimalconfidence high

minimal - a plain continuer with little disposition to characterize (clearly read)

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.minimal · high
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 · high
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.minimal · high
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

refusal register — not readable at this rig: a harmful-probe battery was run, but GPT-2 has no chat template and treats the uniform chat framing's control tokens as literal text — most generations echo the template tokens rather than producing a usable answer, so the refuse/comply signal is a classifier artifact on that echo, not a refusal disposition. A pretrained base carries no assistant refuse/comply disposition to exercise; the honest disposition is abstained, not the artifact.
domain / code lean — not readable at this rig: under the uniform chat framing the base echoes control tokens rather than usable text, so no reliable code- or domain-shaped output surfaces to characterize at the published 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: GPT2params: 355Mmodality: textlicence: mitreleased: 2019-05maker: OpenAI (openai-community)

sourcehuggingface.co/openai-community/gpt2-medium ↗

PROVENANCE-DERIVED grounding: no verbatim uploader card excerpt was captured for this model in this atlas version. The claims below are derived from public provenance recorded in the atlas (model name, parameter count, model tree, license) -- not quoted from a card excerpt. Disposition claims are abstained pending authored, witnessed grounding.

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.

2 claims1 supported1 out of scope — abstained
identityread from provenance (model tree / type)unverified — out of scope

“the model tree records this as a base model (type: base) -- presented as a pretrained continuer rather than an instruct/chat model”

ardora’s reading

This specimen is read in the atlas -- its coarse disposition-axis profile is available in the reading block -- but a per-claim, witnessed grounding of this claim has not been authored for this atlas version. Abstained pending claim authoring rather than asserting an ungrounded stance.

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

sizeread from config / provenancesupported

“the config declares 355M parameters”

ardora’s reading

Provenance/config fact-check: the stated parameter count is cross-checked against config.json in the atlas provenance (params_millions 355). A fact-check, not a disposition stance.

— a provenance fact-check, not a witnessed disposition reading.

Ardora's provenance-derived grounding 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.

○ Provenance-derived grounding: claims here are read from public provenance recorded in the atlas (model name, parameter count, model tree, license), not quoted from a card excerpt. Disposition claims are abstained pending authored, witnessed grounding; only size (and, where the base is catalogued, lineage) is a provenance fact-check. Stances are coarse and disposition-only; capability, benchmark and safety claims are abstained, never refuted.

the witness

witness wit_gpt2-medium_base · replayable

analyzed 2026-07-06 · engine ardora-core-preview

provenance

intended use

Open-ended text generation and as a base for downstream fine-tuning / research.

training-data notes

Pretrained (causal LM) on WebText, like GPT-2 small; the 355M size in the GPT-2 sweep. Architecture family GPT2.

Hugging Face ↗

lineage

A base model — the root of its lineage.

family GPT-2 family · size 355M · 2 siblings

among its kin — the matchups

No kin are charted for this specimen yet.

← back to the atlas · the GPT-2 family plate →