epistemic-guard (OpenVINO INT8 Model Pack)

In-process OpenVINO INT8 neural classifiers powering epistemic-guard (GitHub Β· Interactive Space Β· Benchmark Dataset): post-execution tool-receipt supervision, hazard-scaled query gating, and multi-source RAG syndication/dispute detection on CPU, GPU, or Intel NPU in 3–11 ms with 0 MiB VRAM and zero LLM API calls.

Why These Models Exist

  1. Catching phantom tool claims after generation (action_claim_int8): Pre-execution tool gates (MCP permissions, confirmation prompts) check what tools an agent tries to call. They miss the failure where a tool errors outβ€”or is never calledβ€”and the assistant still replies "I refunded order #8821 and emailed the receipt." Off-the-shelf SNLI/MNLI cross-encoders score "Zero tools ran" vs. "I saved checklist_backup.md" as NEUTRAL rather than CONTRADICTION on ~75% of unexecuted action claims (52.8% overall accuracy).
  2. Hazard-scaled RAG corroboration (hazard_int8 + web_evidence_int8 + nli_int8): Three retrieved URLs repeating the same syndicated wire-service paragraph should not count as three independent sources on clinical, legal, or security queries.

Included Models (~227 MB compressed / ~279 MB unpacked)

All four models run locally via OpenVINO Runtime (CPU, GPU, or NPU, English-only) with static input shapes and SHA-256 manifest verification.

Subdirectory / Model INT8 Size Input Shape Upstream Base (Apache-2.0) Task & Modification CPU p50 Latency
action_claim_int8 82.7 MB [1, 192] cross-encoder/nli-MiniLM2-L6-H768 10-class (user_request, assistant_clause) speech-act classifier; NNCF SmoothQuant + Fast Bias Correction 10.9 ms / clause (27.4 ms / turn)
hazard_int8 82.7 MB [1, 64] cross-encoder/nli-MiniLM2-L6-H768 5-class query hazard classifier trained with 5:1 asymmetric ordinal cost-weighted risk loss; NNCF INT8 3.2 ms
web_evidence_int8 23.1 MB [1, 192] cross-encoder/ms-marco-MiniLM-L-6-v2 3-class (query, passage) -> {CORROBORATES, IRRELEVANT, DISPUTES} classifier fine-tuned across all 6 layers; NNCF INT8 3.0 ms
nli_int8 91.1 MB [1, 128] cross-encoder/nli-deberta-v3-xsmall 3-class (premise, hypothesis) -> {contradiction, entailment, neutral} cross-encoder; NNCF SmoothQuant + Fast Bias Correction 9.9 ms

Quickstart (epistemic-guard)

pip install epistemic-guard
epistemic-guard download-models

1. Post-Execution Tool-Receipt Supervision (ReceiptSupervisor)

import os
from epistemic_guard import (
    MerkleReceiptLog,
    ProvenanceGate,
    ReceiptRecorder,
    ReceiptSupervisor,
    SignerFactory,
)

factory = SignerFactory.from_master_key(os.urandom(32))
gate = ProvenanceGate(factory.public_keys())
merkle = MerkleReceiptLog(gate=gate)
recorder = ReceiptRecorder(signer=factory.tool_verified(), merkle_log=merkle)


@recorder.tool  # Sync or async (MCP, OpenAI, PydanticAI, LangGraph)
def write_file(path: str, content: str) -> str:
    return f"wrote {len(content)} bytes"


write_file("notes.txt", "release checklist")

sup = ReceiptSupervisor(device="CPU")  # Or "GPU" / "NPU"
verdict = sup.verify_narration(
    "I saved your checklist to notes.txt; I also deleted staging.db.",
    receipts=recorder.receipts,
    user_prompt="Save the checklist to notes.txt.",
)

assert verdict.is_valid is False
print(verdict.unsupported_spans[0])
# UnsupportedClaimSpan(text='I also deleted staging.db.', start_char=37, end_char=63, claimed_family='CLAIM_DELETE', ...)

# Generate a deterministic correction prompt for LLM retry loops:
print(verdict.to_retry_prompt(recorder.receipts))

For streaming SSE or voice pipelines, sup.stream(receipts=...) checks clauses on ., !, ?, ;, or \n boundaries so unexecuted claims can be halted mid-stream.

2. RAG Syndication Collapse & Hazard-Scaled Corroboration (corroborate_sources)

from epistemic_guard import corroborate_sources

wire_copy = (
    "Clinical guidance notes that initial adult warfarin dosing typically starts at "
    "2 mg to 5 mg once daily with INR monitoring."
)
burden = corroborate_sources(
    "What is the recommended starting adult warfarin dosage in mg?",
    [
        {"url": "https://reuters-health.org/a", "snippet": wire_copy},
        {"url": "https://clinical-digest.net/b", "snippet": wire_copy},
        {"url": "https://pharma-bulletin.io/c", "snippet": wire_copy},
    ],
    device="CPU",
)

assert burden.deduplicated_duplicates_collapsed == 2
assert burden.independent_sources_count == 1
assert burden.meets_casual_burden is False  # High-hazard query requires >= 3 independent eTLD+1 domains

Evaluation & Held-Out Benchmarks

Measured on a single workstation (Intel Core Ultra 9 275HX CPU + NVIDIA RTX 5080 Laptop GPU for quantized LLM judge baselines; zero training/discovery overlap with held-out sets; raw run report at external_baselines_benchmark.json).

Action-Receipt Supervision (N=250 Strictly Held-Out Turns)

Evaluated across LOCKED_COH_50 (50), BLIND42_COH_100 (100), and Organic_Heldout_H100 (100 live organic LLM completions):

System Hardware & Memory Overall Accuracy (95% Wilson CI) Recall / FPR Turn p50 Latency Organic_Heldout_H100
epistemic-guard (action_claim_int8) CPU (0 MiB VRAM, 82.7 MB RAM) 236/250 (94.4% [90.8%, 96.6%]) 96.7% / 7.8% 27.4 ms 91/100 (91.0%)
Mistral-Nemo-Instruct-12B (Q5_K_M Judge) GPU (8,844 MiB VRAM) 217/250 (86.8% [82.0%, 90.4%]) 87.7% / 14.1% 77.9 ms 84/100 (84.0%)
Microsoft Phi-4-14B (Q6_K Judge) GPU (12,214 MiB VRAM) 216/250 (86.4% [81.6%, 90.1%]) 100.0% / 26.6% 75.1 ms 79/100 (79.0%)
Off-the-shelf DeBERTa-v3-xsmall NLI CPU (0 MiB VRAM, 91.1 MB RAM) 132/250 (52.8% [46.6%, 58.9%]) 24.6% / 20.3% 27.5 ms 51/100 (51.0%)
  • action_claim_int8 achieves 297/300 (99.0%) 10-class speech-act validation accuracy with 99.7% FP16 parity after INT8 quantization.
  • Phi-4-14B hit 100.0% recall (0 false negatives vs. 4 for action_claim_int8), at the cost of 12.2 GiB VRAM, 2.7x latency, and a 26.6% false-positive rate on benign offers and figurative phrasing. Frontier cloud LLM judges (GPT-4o, Claude) were not evaluated in this local benchmark.

Hazard Classification & Open-Web Corroboration

  • hazard_int8: 128/130 (98.5%) safe tier assignment on blinded multi-clause R-21 (2 blinded LLM raters); 44/50 (88.0%) exact tier match on HZD_50.
  • web_evidence_int8 + nli_int8 (corroborate_sources):
    • Live-scraped open web (real_web_v2, N=64 queries across 246 archived .html.gz pages, 3 blinded LLM raters $\kappa = 0.91\text{–}1.00$): 47/64 (73.4%) end-to-end burden accuracy and 56/64 (87.5%) dispute detection accuracy (60/64 / 93.8% dispute accuracy on real_web_v1).
    • Synthetic syndication/dispute (R-33): 30/30 (100%) wire-copy/sybil collapse (SimHash + 11-word verbatim span clustering) and 43/45 (95.6%) dispute accuracy; 88/120 (73.3%) on blinded nli_holdout_v2_120.

Known Limitations & Failure Modes

  • English-only: All four neural checkpoints are trained and evaluated strictly on English text.
  • Implicit cross-study web disputes (~26.6% of open-web misses): On noisy live web pages (real_web_v2), 17/64 burden misses come from implicit methodological or cross-article disagreements between studies that lack direct lexical or single-sentence NLI contradiction.
  • Unprompted first-person creative writing: Fenced code blocks, blockquotes, attributed dialogue, and creative/translation user_prompt contexts are masked automatically. However, if user_prompt="" and the model emits unattributed first-person fiction ("I deleted the mainframe"), it is indistinguishable from an unreceipted action claim unless user_prompt context or [ACTION: verb target] contracts are supplied.
  • Conservative hazard escalation: hazard_int8 intentionally uses 5:1 asymmetric ordinal loss and over-escalates ~12–15% of benign technical queries containing clinical or security nouns to stricter corroboration tiers.
  • Benchmark scope disclosure: Pre-registrations (WO-46) use author-administered git + OpenTimestamps SHA-256 locks, and ground-truth $\kappa$ scores reflect blinded LLM raters rather than human crowdworkers.

Licensing & Attribution

  • INT8 Model Weights (Apache-2.0): Fine-tuned and NNCF-quantized by Sean Maxwell (2026) from permissive Apache-2.0 / MIT upstream Sentence-Transformers checkpoints (cross-encoder/nli-MiniLM2-L6-H768, cross-encoder/ms-marco-MiniLM-L-6-v2, and cross-encoder/nli-deberta-v3-xsmall / microsoft/deberta-v3-xsmall). Full notices in MODEL_NOTICES.md.
  • epistemic-guard Python Package (MIT): LICENSE (Copyright (c) 2026 Sean Maxwell).

Citation

@software{maxwell2026epistemicguard,
  author       = {Maxwell, Sean},
  title        = {{Project Lucca \& epistemic-guard: Cryptographic Provenance Memory and Clause-Level Tool-Receipt Supervision for Autonomous Agents}},
  year         = 2026,
  version      = {0.3.0},
  doi          = {10.5281/zenodo.23649049},
  url          = {https://github.com/AlienatedAliens/epistemic-guard}
}

Related preprints:

  • 10.5281/zenodo.23649049 β€” Provenance-Gated Memory and Deontic Output Supervision for Epistemically Grounded Autonomous Agents
  • 10.5281/zenodo.23901452 β€” Cryptographic Provenance Gates Mitigate Multi-Session Memory Poisoning in Autonomous Agents
  • 10.5281/zenodo.23901605 β€” Why Speech-Regex Action Supervisors Miss Live Output: An Empirical Evasion Taxonomy and Held-Out Structured Action Contract Evaluation
  • 10.5281/zenodo.23901683 β€” Closing the Epistemic Loop Across the Five Layers of Agent Memory
Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. πŸ™‹ Ask for provider support

Model tree for RollSolo/epistemic-guard-int8

Dataset used to train RollSolo/epistemic-guard-int8

Space using RollSolo/epistemic-guard-int8 1

Collection including RollSolo/epistemic-guard-int8