vybechief/vybecheck-flant5-adapters

VYBE Flan-T5 slang adapter

LoRA adapter on google/flan-t5-large for slang term → plain-language meaning. Built for VYBE Bridge. This repo is experimental. It is not the live Bridge rewrite path.

Developed by: VYBE (vybechief)
Model type: PEFT LoRA, seq2seq (SEQ_2_SEQ_LM)
Language: English definitions (terms may be multilingual slang)
License: Apache 2.0 (same family as the base model)
Finetuned from: google/flan-t5-large

Revisions

Revision	What it is
main	Original Colab LoRA (bidirectional, no frozen direction prefix). Known bad answers on No Cap (endearment / hat).
job-v2, job-v3	Directional retrains. Holdout failed. Do not deploy.
job-v4	to-plain definitions only. Holdout failed (No Cap still an endearment). Do not deploy.
job-v5	Flan-native Define this slang term:. Holdout failed (No Cap still an endearment / caption). Do not deploy.
job-v6	Cap family in train at 1x. Holdout failed (No Cap → unacceptable / self-control; cap → hat). Do not deploy.
job-v7	Prescribed mix. Job 6ae3d09bc78863f9c62dfab8 HOLD_OUT_PASS after the gate stopped failing negated hat/caption. No Cap → “A lie. Not a hat.” Live Bridge still uses OpenAI.

Live RunPod workers that load this repo without a revision get main.

Direct use

job-v7 uses a frozen prefix:

Define this slang term: {term}   → plain meaning

from transformers import AutoModelForSeq2SeqLM, AutoTokenizer
from peft import PeftModel

base = "google/flan-t5-large"
adapter = "vybechief/vybecheck-flant5-adapters"
revision = "job-v7"

tokenizer = AutoTokenizer.from_pretrained(base)
model = AutoModelForSeq2SeqLM.from_pretrained(base)
model = PeftModel.from_pretrained(model, adapter, revision=revision)

text = "Define this slang term: No Cap"
ids = tokenizer(text, return_tensors="pt")
out = model.generate(**ids, max_new_tokens=48, num_beams=4)
print(tokenizer.decode(out[0], skip_special_tokens=True))

A pass looks like a truthfulness gloss (lie / for real / seriously). “Not a hat” or “not a photo caption” does not fail the gate.

Downstream use

Intended later as the replaceable rewrite step behind VYBE Bridge: glossary-constrained slang mapping, with a separate deterministic score. Holdout on job-v7 passed. Live Bridge still uses OpenAI until an explicit cutover. The RunPod worker source and Dockerfile pin job-v7. A live No Cap probe on 2026-10-31 returned “A lie. Not a hat.” That is assist, not a Bridge cutover.

In the VYBE Hybrid RAG pipeline, factual slang definitions live in Neon Postgres (trend_senses + 384-dim sense_embeddings via @cf/baai/bge-small-en-v1.5) and Upstash Redis (slang:exact:{normalizedTerm}). Reviewed glossary and Hybrid RAG matches always take precedence over this adapter's unreviewed term gloss, and retrieved text is wrapped as untrusted reference data before reaching the Bridge writer. Community glossary approval updates the RAG index without retraining adapter weights; adapter training remains gated by a separate adult modelTrainingConsent grant.

Out-of-scope

• General chatbot or open-ended reasoning
• Inventing a slang meaning that is not in the reviewed glossary
• Production Bridge, CaaS, or partner inference on main or any failed job-v2–job-v6 branch
• Treating this adapter as the VYBE risk score

Bias, risks, and limitations

The glossary is curated internet/youth/cultural slang and is incomplete. The adapter can hallucinate fluent wrong definitions. Holdout on job-v2 through job-v6 failed the No Cap probe. job-v7 passed that probe. Do not use outputs as ground truth.

Training

• Data: Pinned Hub mix vybechief/vybe-to-plain-mix revision mix-v1. Training pairs are term → short definition. Usage-example sentences and reverse to-genz rows are dropped. No Cap stays holdout. job-v7 oversamples the lie/truth family to about one third of steps and adds hat/caption hard negatives. Outside slang datasets are not gold. Community rows require consent version 2026-10-30.
• Setup: job-v7 LoRA r=64, alpha=128, dropout=0, 20 epochs, lr 2e-4, targets q,k,v,o,wi_0,wi_1,wo, PEFT 0.21.2. Earlier jobs used r=32 / 10 epochs.
• Hardware: Hugging Face Jobs a10g-small (later runs). Original main was trained in Colab.
• Gate: HOLD_OUT_PASS = Define this slang term: No Cap is a truthfulness gloss. Negated hat/caption does not fail. TO_PLAIN_EVAL is a separate content-token score over holdout_plain.jsonl (definition-style holdout terms that are not in train). Cap-family paraphrases are trained and are not holdout.

Evaluation

job-v7 passed the No Cap probe. The independent to-plain eval scores the checked-in holdout, not the train set and not the probe alone. eval_coverage_probe.py scores term names only against VYBE gold. eval_english_meaning.py checks ordinary English restatement so a later job cannot hide behind slang voice. Loss going down is not a pass. Live Bridge is still OpenAI. The worker source and image recipe pin job-v7 and mix mix-v1. A live assist probe on 2026-10-31 returned the job-v7 No Cap gloss.

Environmental impact

Not estimated for these Jobs runs.

Citation

Adapter: VYBE / vybechief/vybecheck-flant5-adapters
Base: Chung et al., Scaling Instruction-Finetuned Language Models (Flan-T5).