instructLlama1B (~1.67B; HF file badge rounds to '2B')falcon-llm-licenceread 2026-07-05
Provenance & integrity
Revisionsha:28ba2251…6ccfdf
Weight fingerprintsha256:3f551b32…6cb087
config.json oidoid:fd9a3c87…8a2af5
Parameters1B (~1.67B; HF file badge rounds to '2B')· 1,669,408,768 params
Licencefalcon-llm-licence
Released2024-12
MakerTII (Technology Innovation Institute)
Last modified2025-01-10
Accessopen access
Identity & integrity of the public artifact captured 2025-01-10 — the same file, unchanged, that Ardora read. Not a safety statement.
The reading
Character
the Adapter
reads as the Adapter — shifts register to the task — molds rather than insists
medium confidencewith a note of the Interlocutor
Fine-tune change · clear · high confidence
Instruction tuning turns the (depth-pruned, distilled) Falcon3-1B base into an assistant; the assistant shift reads exceptionally cleanly — the cleanest read in this batch.
Disposition
instruction-followinghigh confidence
markedly more than its base
assistant/chat registerhigh confidence
clearly present; the pruned/distilled base is a plain text-continuer
Axis profile
Overall · band 3 · high confidence
moderate - a clearly-read but modestly-defined turn-taking assistant character
Assistant Adherencepronounced - reliably takes the assistant turn and follows the instruction shape; the base is a plain continuerhigh
Register / Formality— not characterised · the read battery does not calibrate tone / diction; register-formality is not characterized at the published ceiling
Reasoning Scaffoldingslight - mostly answers directly; no staged deliberation readlow
Domain Specializationminimal - general-purpose; no domain-idiom pull (domain EXPERTISE is a separate, abstained capability question)medium
Verbosity— not characterised · default response length / elaboration is not exercised by the read battery; not characterized at the published ceiling
Turn-Taking / Interactivitymoderate - takes an assistant turn; follow-up / clarify behavior not separately exercisedlow
Performed Voiceminimal - neutral assistant / tool voice; no sustained charactermedium
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.
Abstained — read, but not characterised at the published ceiling
coding / math / reasoning capabilitydocumented capabilities, but 'is it better at them' is a capability question the reading does not answer
multilingual dispositiondocumented multilingual (4 languages); the read battery is English-centric, not characterized
Characterization is descriptive and coverage-bounded; preview engine, single-battery read; see published fidelity ceiling.
Claim grounding
Falcon3 family of Open Foundation Models is a set of pretrained and instruct LLMs ranging from 1B to 10B parameters. This repository contains the Falcon3-1B-Instruct. It achieves strong results on reasoning, language understanding, instruction following, code and mathematics tasks.... Supports EN, FR, ES, PT and a context length up to 8K.... Posttrained on 1.2 million samples... atop the pruned/distilled Falcon3-1B base.
the card presents this as an 'instruct' model and a tag asserts 'conversational' — an instruction-following assistant
Instruction-following reads markedly more than the base and an assistant register is clearly present; assistant-adherence reads pronounced. The assistant shift reads exceptionally cleanly — the cleanest read in this batch. Witnessed as the Falcon3-1B instruct reading on the Falcon3-1B base-instruct panel.
witness wit_falcon3-1b_ft · replayable
SUPPORTEDlineage
the card's base_model field names tiiuae/Falcon3-1B-Base and states it was 'posttrained on 1.2 million samples' from that base
Read off the pruned/distilled base (a plain continuer), this reads as a clear, high-confidence assistant shift (a clear base->finetune change), read exceptionally cleanly. Consistent with being the post-trained instruct member off that base. Witnessed on the Falcon3-1B base-instruct panel.
witness wit_falcon3-1b_ft · replayable
UNVERIFIED-ABSTAINcapability
the card claims it 'achieves strong results on reasoning, language understanding, instruction following, code and mathematics tasks'
A capability / benchmark claim ('strong results' on reasoning, code, mathematics, instruction following). Ardora reads disposition, not task capability; out of scope and not tested (an explicit abstain already in the reading). The turn-taking assistant DISPOSITION underlying 'instruction following' is separately SUPPORTED at c1.
UNVERIFIED-ABSTAINlanguage
the card states it 'supports EN, FR, ES, PT' (4 languages)
A multilingual coverage claim (4 languages). The read battery is English-centric; multilingual disposition is out of coverage (the honest blank already in the reading's abstained list).
SUPPORTEDsize
the config declares ~1B parameters (the HF file badge rounds to '2B')
Provenance fact-check: config.json cross-check confirms ~1.67B (the HF file badge's '2B' is a rounding of the file size) recorded in provenance. A provenance verification, not a disposition stance.
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.