BaseGemma (v1) familyARDORA READ · 2026-07-05

gemma-2b

Google's Gemma v1 2B base model (ungated unsloth mirror) for text generation and fine-tuning.

unsloth/gemma-2b

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: 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.

Characterization is descriptive and coverage-bounded; preview engine, single-battery read; see published fidelity ceiling.

what stayed in shadow — the abstains

domain expertise (medical/legal) — insufficient signal to characterize at the published ceiling
multilingual disposition — the read battery is English-centric; not characterized 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

unsloth/gemma-2b (Gemma v1 2B)

type: basearch: Gemmaparams: 2.5B (~2.51B)modality: textlicence: gemmareleased: 2024-02maker: Google (via unsloth mirror)
tagstext-generationtransformerssafetensorsengemmagemma-2bunslothtext-generation-inference

[Unsloth mirror front-matter:] 'Finetune Mistral, Gemma, Llama 2-5x faster with 70% less memory via Unsloth!' [...] All notebooks are beginner friendly. [The weights are an ungated unsloth mirror of Google's Gemma v1 2B base model.]

sourcehuggingface.co/unsloth/gemma-2b ↗

Public HuggingFace model card (an ungated unsloth mirror of google/gemma-2b) + config.json, quoted as the uploader's self-report as of the capture date. The unsloth card is thin — it leads with the Unsloth finetuning banner over Google's Gemma weights and carries no standalone base-model description. NB (provenance): the unsloth repo is HF-tagged 'apache-2.0', but the weights are Google's Gemma and are governed by the Gemma Terms of Use (upstream google/gemma-2b is 'gemma'-licensed and gated). The HF 'params' chip shows ~3B (a coarse rounding of ~2.51B). Claims below are extracted from THIS snapshot.

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.

3 claims3 supported
identityread from model_namesupported

“the name 'gemma-2b' (no '-it' suffix) marks this as a base/pretrained model”

ardora’s reading

The reading characterizes it as a plain text-continuer with no assistant register — assistant-adherence reads minimal, exactly the disposition of a pretrained base. The base identity is what the reading finds.

witness wit_gemma-2b_base · replayable
lineageread from card_bodysupported

“the card presents it as Google's Gemma v1 2B base, with an instruction-tuned '-it' sibling”

ardora’s reading

This base anchors the base/instruct pair: its instruction-tuned '-it' sibling reads a clear assistant shift off this base, and this base reads as the plain continuer it diverges from. The base-of-pair relationship reads clearly.

witness wit_gemma-2b_base · replayable
sizeread from model_namesupported

“the name implies ~2B parameters (the HF params chip displays ~3B)”

ardora’s reading

Provenance/config fact-check: config declares ~2.51B parameters (params_millions 2506 in the atlas provenance), consistent with the 'Gemma 2B' name; HF's coarse 'params' chip rounds this to ~3B. A fact-check, not a disposition stance.

disposition onlyA provenance cross-check of the stated size, not a disposition reading. The name-implied '2B' and config ~2.51B agree; the HF ~3B chip is a display rounding.

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

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_gemma-2b_base · replayable

analyzed 2026-07-05 · engine ardora-core-preview

provenance

intended use

Base/pretrained model for text generation and downstream fine-tuning.

training-data notes

Google Gemma v1 base; pretrained on ~2T tokens of primarily-English web, code, and math (Gemma family = Gemma architecture, 18 layers, hidden 2048). Ungated unsloth mirror of google/gemma-2b. NOTE: the unsloth mirror repo is HF-tagged 'apache-2.0', but the weights are Google's Gemma and are governed by the Gemma Terms of Use (upstream google/gemma-2b is 'gemma'-licensed and gated). Not present in local cache.

Hugging Face ↗

lineage

A base model — the root of its lineage.

family Gemma (v1) family · size 2B · 1 sibling

← back to the atlas · the Gemma (v1) family plate →