Proposal Explorer

Add claims to see which ones the mechanisms address.

Operational use

• Operational use Sampled inference recomputation for checking untrusted workers' activations against the declared model, prompt and precision

Built for an adversarial prover

Sampled inference recomputation

No new hardware needed

Sampled inference recomputation

Attack testing records published testing for this use. It does not by itself show independent review, a formal proof or that a deployed system is secure.

• Sampled inference recomputation: Analysis
• Bandwidth limits and compartmentalization, RAND secure inference data center (SIDC) design: Analysis

Family findings

• Bandwidth limits and compartmentalization

  Context for RAND secure inference data center (SIDC) design. Findings from the mechanism family appear here as context. They apply to an implementation only when its own record lists them, under the conditions stated there.

  • Low-communication training reduces the bandwidth training needs in Bandwidth limits and compartmentalization

    DiLoCo matched fully synchronous training on 8 workers while communicating 500 times less. Rahman writes that this family of methods theoretically allows large-scale training with less than 100 Mbps. Lucid includes these methods in its bounds, but notes that extreme activation compression, architectures with unusually small inter-layer widths, or modular paradigms could erode the margin. ^{16 17 18}

    Open question · Theoretical argument. On the record

  • Operator control of pod routing collapses the bound in Bandwidth limits and compartmentalization

    Lucid's analysis finds that if the operator can freely assign pods to routers, it could dedicate a whole cell of 100 or more pods to one pipeline stage. The bound then falls to about 90–220x uncompressed and as low as about 25x with compression. The proposed mitigation, auditor-controlled random assignment that is periodically re-randomized, has not been implemented. ¹⁸

    Known failure · Theoretical argument · Significant · Open. On the record

  • Undeclared local storage raises per-pod capacity in Bandwidth limits and compartmentalization

    More memory or storage per pod helps an adversary. Lucid requires per-pod storage to be declared, capped and physically inspected. ¹⁸

    Scope limitation · Theoretical argument. On the record

  • Training within one pod is not covered in Bandwidth limits and compartmentalization

    Lucid's bounds concern pre-training models larger than the pods are sized for. Training models that fit in one pod, fine-tuning and reinforcement-learning post-training within one pod are outside the modelled threat. ¹⁸

    Scope limitation · Open question. On the record

  • Parallel scale-up switches are hard enforcement points in Bandwidth limits and compartmentalization

    In GB200 topologies, GPUs reach GPUs in other nodes through NVSwitches without a NIC on the path. Amodo notes that limits are hard to enforce there because many switches work in parallel, so compromising one or two would bypass the limit. ¹⁹

    Known failure · Theoretical argument · Significant · Open. On the record

Open significant failures

2 failures in 1 mechanism

• Tolerance for numerical noise leaves a covert channel in Sampled inference recomputation

  Schemes that accept approximate matches can put an upper bound on an adversary's covert bandwidth, but they cannot close the channel. The weight-exfiltration detector cut exfiltratable information to under 0.5%, not to zero, on a 30-billion-parameter mixture-of-experts model under benign prompt traffic. Its authors called the channel's size under adversarial prompts an open empirical question. An independent study showed that an adversary who controls the prompts roughly doubles the bits leaked per token. Across six models, that cut the slowdown from 146–254 times under benign prompts to 60–118 times. The attack widens the exfiltration bound. It does not target the check that outputs match the declared model. ^{1 9 12}

  Known failure · Demonstrated attack · Significant · Open. On the record

  Related mechanism Operational use Deterministic and bit-exact inference: Bit-exact inference would remove the numerical tolerance if exact replay can be deployed with the required weights and configuration. A pointer, not evidence that this failure is mitigated. Add

• Some inference optimizations are not covered in Sampled inference recomputation

  TOPLOC's authors state that it cannot detect speculative decoding in which a cheaper model does the decoding. They did not test whether it distinguishes types of key-value (KV) cache compression. DiFR was evaluated only on sampling from a single model. Its authors sketch an extension to one speculative-decoding algorithm but do not test it. ^{2 3}

  Known failure · Theoretical argument · Significant · Open. On the record

Scope limitations

• Only recorded traffic is checked in Sampled inference recomputation

  Recomputation checks that recorded, declared workloads are correct. It cannot show that the record is complete. The published schemes do not cover hidden workloads run on the same compute, or substituted work. Rinberg et al. say their exfiltration-detection scheme cannot stand alone. ^{1 13}

  Scope limitation · Theoretical argument. On the record

  Related mechanism Proposed Network taps and certifiers: Taps copy and hash all traffic on the monitored links, which bears on whether the traffic record is complete. They do not show what else ran on the same chips. A pointer, not evidence that this failure is mitigated. Add

• Everything rests on the trusted setup in RAND secure inference data center (SIDC) design

  Reference measurements for model weights and reference data are established in a trusted setup phase. The report states that the system cannot detect compromise that happened before ingestion if the trusted setup itself is compromised. ¹⁵

  Scope limitation · Theoretical argument. On the record

• Security weakens over long operation in RAND secure inference data center (SIDC) design

  The authors claim that the facility can withstand attacks at the OC5 level for a five-year operational period. They expect its ability to withstand long OC5 campaigns to become less robust the longer the facility remains in operation. ¹⁵

  Scope limitation · Theoretical argument. On the record

Open minor failures

1 mechanism with minor failures

• Mixed hardware widens the honest baseline in Sampled inference recomputation

  When honest reference runs span different GPU types, the spread of benign scores grows. In DiFR's tests on Qwen3-30B-A3B, pooling A100 and H200 runs left Token-DiFR unable to separate the two smallest tested changes, a temperature of 1.1 instead of 1.0 and a simulated top-2 sampling bug, at the target false-positive rate, while cross-entropy separated them. Matched provider and verifier environments, or pooling that weights rare large deviations, restored detection. ²

  Known failure · Open question · Minor · Open. On the record

Not yet demonstrated

Proposed Bandwidth limits and compartmentalization

Mechanisms on the map that are not in the proposal. Pointers, not recommendations. ?

• Operational use Deterministic and bit-exact inference Add
  • Bears on the open significant failure “Tolerance for numerical noise leaves a covert channel” in Sampled inference recomputation. Bit-exact inference would remove the numerical tolerance if exact replay can be deployed with the required weights and configuration.
• Proposed Network taps and certifiers Add
  • Sampled inference recomputation waits on it. In tap-based retrofit designs, recording all inference traffic needs network taps and recomputation servers that can ingest it, in the worst case one recomputation-server network interface per inference front-end interface.

Blockers

9 blockers recorded

• Sampled inference recomputation
  • In tap-based retrofit designs, recording all inference traffic needs network taps and recomputation servers that can ingest it, in the worst case one recomputation-server network interface per inference front-end interface. Coverage & hidden compute. Waits on Network taps and certifiers ^{8 20}
  • In retrofit designs, the recomputation server must sit inside the prover's data centre, possibly under the prover's physical control, and still be protected from a compromised provider, which Amodo rates 'not on track'. Hardware trust ^{1 8}
  • No independent red-team of a recomputation consistency check has been published (the one independent attack study targets the weight-exfiltration bound), and Amodo rates recomputation red-teaming 'not started'. Adversarial validation ^{8 9}
  • Tolerance-based checks need calibration on trusted hardware and exact knowledge of the provider's sampling procedure, and in one prototype a sampling-implementation mismatch produced large spurious differences. Performance & compatibility ^{2 7}
  • The verifier needs the model weights, so checking a closed-weights model requires a trusted, confidential recomputation environment, which the retrofit designs place inside the prover's facility. Privacy & leakage ^{2 13 14}
• Bandwidth limits and compartmentalization
  • No prototype exists; RAND recommends prototyping key security features and integration now. Adversarial validation ¹⁵
  • The report describes internal integrity checks, audit logging and accreditation, but no way for a party outside the operator to verify the facility's properties. Access & governance ¹⁵
  • Human review of every prompt and response makes each request take three to five minutes, with the review steps as the rate-limiting factor. Performance & compatibility ¹⁵
  • Detailed design information is withheld from the public report and is to be evaluated privately with stakeholders, which limits independent public scrutiny. Access & governance ¹⁵

From the family or selected implementation's record. ?

Model weights

Depends on the design for Sampled inference recomputation.

Unspecified for Bandwidth limits and compartmentalization. Check the implementation record.

Inputs and outputs

Depends on the design for Sampled inference recomputation.

Unspecified for Bandwidth limits and compartmentalization. Check the implementation record.

Training data

Not involved: Sampled inference recomputation.

Unspecified for Bandwidth limits and compartmentalization. Check the implementation record.

Sampled inference recomputation

• Proposed AI 2040 inference-only verification stack Proposed architecture, AI Futures Project
• Research demo DiFR (Divergence From Reference) Research prototype
• Proposed Low-trust AI compute verification system overview Proposed architecture, Machine Intelligence Research Institute
• Proposed SASH confidential network logger Research prototype, Singapore AI Safety Hub (SASH)
• Operational use TOPLOC Open-source project, Prime Intellect

Bandwidth limits and compartmentalization

• Proposed AI 2040 inference-only verification stack Proposed architecture, AI Futures Project
• Proposed RAND secure inference data center (SIDC) design Proposed architecture, RAND

1. Verifying LLM Inference to Detect Model Weight Exfiltration, R. Rinberg et al. (2025). Original
2. DiFR: Inference Verification Despite Nondeterminism, A. Karvonen et al. (2025). Original
3. TOPLOC: A Locality Sensitive Hashing Scheme for Trustless Verifiable Inference, J. M. Ong et al. (2025). Original
4. PrimeIntellect-ai/toploc (GitHub repository), Prime Intellect (2025). Original
5. INTELLECT-2: A Reasoning Model Trained Through Globally Decentralized Reinforcement Learning, Prime Intellect Team et al. (2025). Original
6. adamkarvonen/difr (GitHub repository), A. Karvonen (2025). Original
7. Scaling Recomputation Inference Verification, Amodo Design (2026). Original
8. AI 2040 Plan A — Verification SITREP, Amodo Design (2026). Original
9. Adversarial Entropy Inflation Against Gumbel-Based Inference Verification, N. Kezins (2026). Original
10. SYNTHETIC-2 Release: Four Million Collaboratively Generated Reasoning Traces, Prime Intellect (2025). Original
11. An Inference Verification Prototype — Stage 1, Amodo Design (2026). Original
12. Bit-Exact AI Inference Verification Without Performance Tradeoffs, N. Cankaya (2026). Original
13. Example Schemes for Verifying High-Stakes AI Agreements, Amodo Design (2026). Original
14. A System Overview for Near-Term, Low-Trust AI Compute Verification, N. Cankaya (2026). Original
15. Highly Secure Inference Data Centers: A Vertically Integrated Strategy for Security Engineering, S. F. Comer et al. (2026). Original
16. DiLoCo: Distributed Low-Communication Training of Language Models, A. Douillard et al. (2024). Original
17. Does Distributed Training Undermine Compute Governance?, R. Rahman (2026). Original
18. Traffic Shaping for Workload Classification, Lucid Computing (2026). Original
19. The Tray as a Bandwidth Boundary, Amodo Design (2026). Original
20. Verification Plan, R. Dean (2026). Original