metafloor-ai/manifest-specialist-risk-and-resilience-v0.6.0

Manifest Specialist — Risk & Resilience · v0.6.0

A specialized supply-chain domain-expert — focused entirely on risk & resilience.

Manifest Risk & Resilience is a domain-expert in the Manifest family: where the general Manifest orchestrator models cover all of supply chain, this one is tuned specifically for supply-chain risk and resilience — disruption, supplier failure, monitoring, mitigation, and continuity.

• Family: Manifest · Type: domain-expert (specialized) · Size: 2B
• Specialty: supply-chain risk & resilience

Preferred 100.0% of the time over the base model — 85.0% with matched format

On a focused 20-question risk & resilience benchmark, an independent LLM judge panel preferred this model's answer over the base model's 100.0% of the time (20 / 0 / 0 W/L/T), and 85.0% with the same answer format. (20 items is a small sample — treat exact figures as directional.)

What's new in v0.6

• Retrained on the expanded v0.6 supply-chain data — this specialist's focused split now draws from MetaFloor's 32k-example corpus (4k in-domain examples).
• Scored on the current 20-question risk & resilience benchmark; the 35B orchestrator is retired this release and a new 9B tier is coming.
• This is a new versioned repo — the previous v0.5 release (metafloor-ai/manifest-specialist-risk-and-resilience, unchanged) keeps its own history and download stats.

v0.5 → v0.6 (preferred over the untuned base, same benchmark):

Release	Preferred over base	vs base + matched format
v0.5.0	70.0%	72.5%
v0.6.0	100.0%	85.0%

Best for

Risk and resilience questions where a focused, practitioner's answer beats a generic one:

• Disruption response, supplier-failure and continuity planning
• Risk monitoring, early-warning and alerting design
• Resilience trade-offs (buffers, dual-sourcing, network redundancy)

See the difference

Same question. Base model vs Manifest Risk & Resilience.

Ask: "What is the failure mode known as alert fatigue in supply-chain risk monitoring, and how can data analytics mitigate it?"

Base model → "### Failure Mode: Alert Fatigue. Alert fatigue is a critical failure mode where the volume, frequency, or complexity of alerts…" — a generic textbook definition.

Manifest Risk & Resilience → "Alert fatigue is the state where an organization receives too many, unfiltered risk alerts, so it stops paying attention to the real ones and misses the ones it never saw — a failure of signal-to-noise management, not a lack of data…" — sharp, practitioner framing.

The Manifest family

Two kinds of models:

🧭 Orchestrators — general-purpose, handle any supply-chain area

Model	Size	Preferred over base	Status
Manifest 0.8B	0.8B	79.9%	✅ available
Manifest 2B	2B	85.1%	✅ available
Manifest 4B	4B	95.1%	✅ available
Manifest 9B	9B	—	🔜 coming soon

🎯 Domain-experts — specialized for a single area

Model	Preferred over base	Status
Manifest Specialist · Risk & Resilience	100% (20 / 0)	✅ available
Manifest Specialist · Inventory Optimization	80.0%	✅ available
Manifest Specialist · Demand Planning	82.5%	✅ available

Orchestrators are scored on the general supply-chain benchmark; domain-experts on their focused domain benchmark (20 items each — treat those exact figures as directional).

How to use

Manifest Risk & Resilience is a LoRA adapter (~44 MB), applied on top of its base model at load time.

from peft import PeftModel
from transformers import AutoModelForCausalLM, AutoTokenizer

base = "Qwen/Qwen3.5-2B"  # base model — see "Built on" below
tok = AutoTokenizer.from_pretrained(base)
model = AutoModelForCausalLM.from_pretrained(base, device_map="auto")
model = PeftModel.from_pretrained(model, "metafloor-ai/manifest-specialist-risk-and-resilience-v0.6.0")

SYSTEM = "You are a senior supply chain expert. Answer correctly and concisely."
msgs = [{"role": "system", "content": SYSTEM},
        {"role": "user", "content": "A strategic supplier just failed an audit. Walk me through the risk response."}]
inputs = tok.apply_chat_template(msgs, add_generation_prompt=True, return_dict=True, return_tensors="pt")
out = model.generate(**inputs, max_new_tokens=512, temperature=0.7)
print(tok.decode(out[0][inputs["input_ids"].shape[-1]:], skip_special_tokens=True))

Prompt tip: include the asker's role and operating scale (revenue, SKUs, suppliers, nodes, lead time) for the sharpest answers.

Training details

Method	LoRA (PEFT 0.20.0), rank 16 / alpha 16 / dropout 0.05
Target modules	all attention + MLP projections
Trainable params	10,911,744 (~0.9% of the 1.21B base)
Epochs	3
Training examples	~4,000
Final loss	1.36 (from 2.64)

The training data is a focused, proprietary supply-chain risk dataset and is not open-sourced — only the held-out evaluation benchmark (supply-chain-eval) is public.

Evaluation

Scored on a focused 20-question risk benchmark — pairwise LLM-as-judge (2-model panel), Manifest's answer vs the base model's for the same prompt.

Comparison	Win-rate	W / L / T
Manifest Risk & Resilience vs base	70.0%	13 / 5 / 2
Manifest Risk & Resilience vs base + format	72.5%	14 / 5 / 1

Benchmark: supply-chain-eval (risk split). This is an early, focused benchmark — the sample is small, so treat it as directional.

Intended use & limitations

• Intended use: decision-support and drafting for supply-chain risk & resilience questions.
• Out of scope: not legally binding, contractual, or safety-critical guidance; no access to your live systems. Verify outputs before acting.
• Limitations: English-only; specialized to risk (use a Manifest orchestrator for general questions); trained on synthetic data; standard LLM risks apply.

License

Released under CC-BY-NC-4.0 — free for research and non-commercial use, with attribution. Commercial use requires a license from MetaFloor — get in touch at metafloor.ai.

Built on

Manifest Risk & Resilience is a LoRA adapter over Qwen/Qwen3.5-2B (used under its own license); the base model is required to load the adapter.

Citation

@misc{metafloor_manifest_risk_2b,
  title  = {Manifest Risk & Resilience: a supply-chain risk domain-expert (MetaFloor Manifest family)},
  author = {MetaFloor AI},
  year   = {2026},
  howpublished = {\url{https://huggingface.co/metafloor-ai/manifest-specialist-risk-and-resilience-v0.6.0}}
}