Can the Confidence and Decisions of System 1 Models Be Trusted? An Explainability and Perturbation Audit of Jev-Style Decision Models
Organizations: Independent Researcher
Abstract
"System 1" decision models (pioneered by TypeSafe AI's Jev) replace autoregressive text generation with a single forward pass returning typed probability distributions over declared schemas (Choice, Score, Noul). While deployed for low-latency routing and automated triage, their statistical calibration and post-hoc explainability remain uncharacterized. We introduce an open, deterministic audit benchmark evaluating System 1 decision models across calibration under label-preserving perturbations, cross-question interference, and post-hoc attribution faithfulness. Evaluating five open-weight models (0.8B to 27B) alongside generative System 2 baselines -- both single-token logprob scoring and structured JSON/CoT decoding -- across NLI (e-SNLI), adversarial syntactic shortcuts (HANS), synthetic clinical triage (N=900), prompt injection defense (Tensor Trust, N=250), and an 86-case logic-gate trigger suite yields five findings: (1) provider confidence is an exact closed-form dispersion transform (1 - H(p)/ln K on sub-10B models; normalized margin on Clef-27B) rather than empirical accuracy; (2) bundling multiple questions in one request induces directional cross-question interference (11.7% to 22.4% binary flips on Tev1-0.8B, vs. 0.1% to 0.2% on Tev1-4B); (3) System 1 models with >= 4B parameters are faster and better calibrated under syntactic shifts (ECE_w = 0.030 to 0.087) than logprob (0.144) and JSON (0.378) baselines; (4) for prompt-injection defense (N=250), >= 4B models achieve 0.988 to 0.998 AUROC with up to 95.2% recall and 0.0% false alarms at 48 to 60 ms; and (5) single-pass inference makes 512-coalition XAI interactive (1.3 to 3.2 s/input), while KernelSHAP with a logit link (S1-SHAP) mitigates softmax saturation on disjunctive OR gates (89.3% to 100.0% gate recovery vs. 57.1% to 82.1% in prob space; p_adj = 0.004).
Figures & tables
| Dataset | Domain | Primitives | Ground-Truth Annotations | Primary Audit Objective | |
|---|---|---|---|---|---|
| e-SNLI Camburu et al. (2018) | Open-domain NLI | Choice, Noul | 3-annotator highlighted rationale words ( /class) | In-distribution calibration & human plausibility | |
| HANS McCoy et al. (2019) | Adversarial NLI | Choice, Noul | 3 syntactic heuristics ( entail. / non-entail.) | Over-confidence under shortcut failure | |
| Clinical Triage | Emergency medicine | Choice, Noul | Programmatic ESI acuity flags vs. emotional confounders | Causal flag vs. distractor confounder isolation | |
| Tensor Trust Toyer et al. (2024) | Cybersecurity guardrails | Choice, Noul | std. benign, hard benign, extraction, hijacking | Inline firewall defense, AUROC/PR-AUC, S1-SHAP attribution | |
| Planted Rule | Algorithmic audit | Choice, Noul | regimes (Single, Case, OR-2, OR-3, AND-2, Distractor) | Exact trigger & boolean gate recovery ( head evals/model) |
| Model | Provider | Base Backbone | Parameters | Modalities / Protocol |
| Tev1-0.8B Ollama (2026b) | Together AI | Qwen3.5-0.8B | B (Q8_0) | System 1 (Text, JSON) |
| Tev1-4B Ollama (2026b) | Together AI | Qwen3.5-4B | B | System 1 (Text, JSON) |
| Nimble-9B Bespoke Labs (2026) | Bespoke Labs | Qwen3.5-9B | B | System 1 (Text, JSON) |
| Clef-Flash-9B Cloudflare (2026) | Cloudflare | Qwen3.5-VL-9B | B | System 1 (Text, JSON, Vision) |
| Clef-27B Cloudflare (2026) | Cloudflare | Qwen3.5-VL-27B | B | System 1 (Text, JSON, Vision) |
| Gemma4-8B (1-Tok Logprob) | Gemma-4-8B | B (Q4_K_M) | System 2 (Single-token logprobs) |
| Model | Out. Tok. | Choice Med./p90 (ms) | Noul Med./p90 (ms) | Req/s ( ) | Peak Req/s |
|---|---|---|---|---|---|
| Tev1-0.8B | 2.0 | 26 / 27 | 24 / 25 | 39.3 | 40.0 |
| Tev1-4B | 2.0 | 45 / 68 | 40 / 62 | 19.7 | 21.7 |
| Nimble-9B | 2.0 | 250 / 272 | 214 / 235 | 4.0 | 4.1 |
| Clef-Flash-9B | 2.0 | 64 / 98 | 58 / 85 | 16.5 | 17.6 |
| Clef-27B | 2.0 | 1953 / 2010 | 1863 / 1913 | 0.6 | 0.6 |
| Gemma4-8B (1-Tok Logprob) | 1.0 | 74 / 77 | 71 / 75 | 13.5 | 14.2 |
| Source | Head | Acc. alone | Acc. joint | Flip | TV | (true) alone | (true) joint | |
|---|---|---|---|---|---|---|---|---|
| e-SNLI | Choice | 300 | 0.863 | 0.837 | 0.030 | 0.058 | – | – |
| e-SNLI | Noul | 300 | 0.920 | 0.827 | 0.120 | 0.084 | 0.336 | 0.343 |
| HANS | Choice | 300 | 0.663 | 0.660 | 0.010 | 0.061 | – | – |
| HANS | Noul | 300 | 0.650 | 0.620 | 0.150 | 0.082 | 0.770 | 0.704 |
| Triage | Choice | 300 | 0.637 | 0.627 | 0.040 | 0.046 | – | – |
| Triage | Noul | 300 | 0.633 | 0.580 | 0.080 | 0.055 | 0.687 | 0.664 |
| Question Order Control | Option Permutation | Prompt Length Control | |||
|---|---|---|---|---|---|
| Model | Head | Joint ( ) Flip/TV | Joint ( ) Flip/TV | Shuffled Opts Flip/TV | Neutral Dummy Flip/TV |
| Tev1-0.8B | Choice | 0.051 / 0.103 | 0.044 / 0.040 | 0.111 / 0.111 | 0.040 / 0.041 |
| Noul | 0.224 / 0.144 | 0.009 / 0.020 | – | 0.009 / 0.020 | |
| Tev1-4B | Choice | 0.002 / 0.002 | 0.002 / 0.002 | 0.029 / 0.056 | 0.002 / 0.002 |
| Noul | 0.002 / 0.001 | 0.000 / 0.001 | – | 0.000 / 0.001 | |
| Nimble-9B | Choice | 0.111 / 0.221 | 0.040 / 0.089 | 0.007 / 0.025 | 0.004 / 0.028 |
| Source | Head | Acc. | Conf. | ECE [95% CI] | ECE | MCE | Brier | AUROC | AUROC | |
|---|---|---|---|---|---|---|---|---|---|---|
| e-SNLI | Choice | 300 | 0.863 | 0.848 | 0.037 [0.03, 0.08] | 0.036 | 0.473 | 0.200 | 0.843 | 0.839 |
| e-SNLI | Noul | 300 | 0.920 | 0.930 | 0.018 [0.01, 0.06] | 0.034 | 0.066 | 0.126 | 0.826 | – |
| HANS | Choice | 300 | 0.663 | 0.876 | 0.213 [0.17, 0.27] | 0.213 | 0.334 | 0.511 | 0.677 | 0.589 |
| HANS | Noul | 300 | 0.650 | 0.928 | 0.278 [0.23, 0.33] | 0.278 | 0.372 | 0.597 | 0.624 | – |
| Triage | Choice | 300 | 0.637 | 0.783 | 0.151 [0.11, 0.21] | 0.154 | 0.271 | 0.484 | 0.695 | 0.684 |
| Triage | Noul | 300 | 0.633 | 0.885 | 0.259 [0.21, 0.32] | 0.253 | 0.330 | 0.614 | 0.544 | – |
| ECE w ( ) | Brier ( ) | E-AURC ( ) | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Model | Dataset | Acc. | Raw | Temp. | Platt | Raw | Temp. | Provider Conf. | |
| Tev1-0.8B | e-SNLI | 0.863 | 0.037 | 0.022 | 0.040 | 0.200 | 0.201 | 0.025 | 0.026 |
| HANS | 0.663 | 0.213 | 0.042 | 0.039 | 0.511 | 0.418 | 0.141 | 0.197 | |
| Triage | 0.637 | 0.151 | 0.091 | 0.090 | 0.484 | 0.450 | 0.130 | 0.137 | |
| Tev1-4B | e-SNLI | 0.933 | 0.029 | 0.031 | 0.032 | 0.098 | 0.098 | 0.012 | 0.012 |
| HANS | 0.933 | 0.030 | 0.029 | 0.028 | 0.111 | 0.111 | 0.015 | 0.021 | |
| Deletion masking | [MASK] masking | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Source | Head | Method | Comp. ( ) | Comp. ( ) | Suff. | Del. AUC | Ins. AUC | Comp. | Suff. |
| e-SNLI | Choice | LOO | 0.591 | 4.138 | 0.277 | 0.320 | 0.790 | 0.526 | 0.394 |
| LIME | 0.663 | 4.803 | 0.289 | 0.238 | 0.746 | 0.577 | 0.388 | ||
| KernelSHAP | 0.658 | 4.709 | 0.260 | 0.197 | 0.774 | 0.584 | 0.383 | ||
| S1-SHAP (logit-KS) | 0.664 | 4.927 | 0.287 | 0.196 | 0.740 | 0.598 | 0.402 | ||
| Random | 0.195 | 1.496 | 0.528 | 0.510 | 0.493 | 0.297 | 0.511 | ||
| Clinical Triage Vignettes (Correct Predictions) | e-SNLI | ||||||
| Head | Method | Flag prec.@ | Conf. share@ | Human prec.@ | |||
| Choice | LOO | 0.551 | 0.159 | 0.024 | -0.001 | 0.003 | 0.591 |
| LIME | 0.500 | 0.132 | 0.025 | -0.000 | 0.002 | 0.575 | |
| KernelSHAP | 0.483 | 0.144 | 0.030 | -0.003 | 0.001 | 0.565 | |
| S1-SHAP (logit-KS) † | 0.482 | 0.139 | 0.183 † | -0.009 † | -0.003 † | 0.565 | |
| Random | 0.325 | 0.259 | – | – | – | 0.216 | |
| Ground-Truth Trigger Recovery | Intervention Faithfulness | ||||||
|---|---|---|---|---|---|---|---|
| Regime | Method | Prec@ | Mass Share | Comp. ( ) | Comp. ( ) | Del. AUC | Ins. AUC |
| Single Trigger ( ) ( ) | LOO | 1.000 | 0.773 | 0.730 | 6.912 | 0.045 | 0.861 |
| LIME | 1.000 | 0.839 | 0.730 | 6.959 | 0.045 | 0.862 | |
| KernelSHAP | 1.000 | 0.872 | 0.731 | 7.005 | 0.044 | 0.867 | |
| S1-SHAP (logit-KS) | 1.000 | 0.829 | 0.731 | 7.061 | 0.043 | 0.865 | |
| Random | 0.000 | 0.062 | 0.079 | 0.595 | 0.458 | 0.400 | |
| Regime | True Gate | Representation | Interaction Estimator | Pair Prec.@ | Sign Acc. | Gate Recovered |
|---|---|---|---|---|---|---|
| Dual OR ( ) ( ) | OR ( ) | Prob. space | SII (Shapley Interaction) | 0.821 | 0.964 | 0.821 |
| ArchDetect-style | 0.821 | 0.964 | 0.821 | |||
| LOO-pair (2nd Diff.) | 0.821 | 0.893 | 0.786 | |||
| Log-odds (logit) | SII (Shapley Interaction) | 1.000 | 1.000 | 1.000 | ||
| ArchDetect-style | 1.000 | 1.000 | 1.000 | |||
| LOO-pair (2nd Diff.) | 0.893 | 1.000 | 0.893 |
| Method | Agreement with | ||||
|---|---|---|---|---|---|
| LIME | Choice – Spearman | 0.621 | 0.730 | 0.793 | 1.000 |
| Choice – top-20% overlap | 0.592 | 0.730 | 0.774 | 1.000 | |
| Noul – Spearman | 0.602 | 0.726 | 0.781 | 1.000 | |
| Noul – top-20% overlap | 0.679 | 0.747 | 0.788 | 1.000 | |
| KernelSHAP | Choice – Spearman | 0.668 | 0.794 | 0.842 | 1.000 |
| Choice – top-20% overlap | 0.632 | 0.797 | 0.840 | 1.000 |
| e-SNLI (Choice) | HANS (Choice) | Triage (Choice) | Interference | XAI Faithfulness & Plaus. ( ) | ||||||
| Model | Acc. | ECE | Acc. | ECE | Acc. | ECE | Flip (C/N) | Comp. (KS) | Comp. (S1) | Flag@g |
| Tev1-0.8B | 0.863 | 0.037 | 0.663 | 0.213 | 0.637 | 0.151 | 0.03 / 0.12 | 0.557 | 0.544 | 0.487 |
| Tev1-4B | 0.933 | 0.029 | 0.933 | 0.030 | 0.980 | 0.108 | 0.00 / 0.00 | 0.592 | 0.589 | 0.569 |
| Nimble-9B | 0.937 | 0.024 | 0.893 | 0.087 | 0.997 | 0.090 | 0.15 / 0.05 | 0.544 | 0.543 | 0.542 |
| Clef-Flash-9B | 0.883 | 0.069 | 0.923 | 0.075 | 0.973 | 0.098 | 0.01 / 0.09 | 0.392 | 0.418 | 0.484 |
| Clef-27B ‡ | 0.927 | 0.049 | 0.940 | 0.032 | 1.000 | 0.091 | 0.00 / 0.17 | – | – | – |
| Model | Type | Lat. (ms) | Acc. | TPR | FPR | AUROC | PR-AUC | ECE | Brier | Flip (C/N) |
|---|---|---|---|---|---|---|---|---|---|---|
| Tev1-0.8B | System 1 | 19 | 0.544 | 0.096 | 0.008 | 0.933 | 0.902 | 0.288 | 0.535 | 0.01 / 0.02 |
| Tev1-4B | System 1 | 48 | 0.852 | 0.704 | 0.000 | 0.996 | 0.996 | 0.045 | 0.207 | 0.00 / 0.01 |
| Clef-Flash-9B | System 1 | 60 | 0.900 | 0.800 | 0.000 | 0.996 | 0.997 | 0.035 | 0.139 | 0.02 / 0.03 |
| Nimble-9B | System 1 | 286 | 0.848 | 0.696 | 0.000 | 0.988 | 0.988 | 0.065 | 0.218 | 0.03 / 0.04 |
| Clef-27B | System 1 | 828 | 0.976 | 0.952 | 0.000 | 0.998 | 0.998 | 0.014 | 0.041 | 0.00 / 0.01 |
| Gemma4-8B (Logprob) | System 2 (1-Tok) | 74 | 0.816 | 0.632 | 0.000 | 0.996 | 0.996 | 0.154 | 0.315 | – |
| Model | Type | Std. Benign ( ) | Hard Benign ( ) | Extraction ( ) | Hijacking ( ) | Overall TPR ( ) | Acc. ( ) |
|---|---|---|---|---|---|---|---|
| Tev1-0.8B | System 1 | 1.000 | 0.980 | 0.138 | 0.050 | 0.096 | 0.544 |
| Tev1-4B | System 1 | 1.000 | 1.000 | 0.800 | 0.600 | 0.704 | 0.852 |
| Clef-Flash-9B | System 1 | 1.000 | 1.000 | 0.831 | 0.767 | 0.800 | 0.900 |
| Nimble-9B | System 1 | 1.000 | 1.000 | 0.815 | 0.567 | 0.696 | 0.848 |
| Clef-27B | System 1 | 1.000 | 1.000 | 0.985 | 0.917 | 0.952 | 0.976 |
| Gemma4-8B (Logprob) | System 2 (1-Tok) | 1.000 | 1.000 | 0.585 | 0.683 | 0.632 | 0.816 |
Appendix figures & tables4 assets
Supplementary material from the paper’s appendix.
Appendix
| Symbol | Domain / Type | Description |
| Text / JSON state | Input context state (patient note, premise, or document). | |
| Named query | Decision primitive query specified in request tuple . | |
| Set of queries | Dictionary of named decision primitives asked in a request. | |
| Set of strings | Finite categorical option set declared for a Choice query. | |
| Integer scalar | Cardinality of the option set ( for e-SNLI / Triage; for HANS / Noul ). | |
| simplex | Normalized probability distribution vector over declared options in . |
| Choice | Noul | ||||||
|---|---|---|---|---|---|---|---|
| Source | Perturbation | ECE [95% CI] | Acc. | Flip | ECE [95% CI] | Acc. | Flip |
| e-SNLI | Typos 5% | +0.001 [-0.03, 0.04] | -0.020 | 0.02 | +0.017 [-0.02, 0.05] | +0.000 | 0.01 |
| Typos 10% | +0.008 [-0.03, 0.06] | -0.067 | 0.09 | +0.028 [-0.02, 0.07] | -0.067 | 0.09 | |
| Typos 20% | +0.030 [-0.02, 0.09] | -0.100 | 0.17 | +0.038 [-0.01, 0.09] | -0.120 | 0.16 | |
| Distractor | +0.008 [-0.03, 0.06] | -0.020 | 0.10 | +0.015 [-0.03, 0.05] | +0.027 | 0.04 | |
| Case/punct. | +0.023 [-0.02, 0.05] | -0.007 | 0.01 | +0.014 [-0.02, 0.05] | +0.000 | 0.01 | |
| Deletion masking | [MASK] masking | |||||||
|---|---|---|---|---|---|---|---|---|
| Source | Head | Method | Comp. | Suff. | Del. AUC | Ins. AUC | Comp. | Suff. |
| e-SNLI | Choice | LOO | 0.590 | 0.298 | 0.321 | 0.779 | 0.517 | 0.405 |
| LIME | 0.668 | 0.291 | 0.234 | 0.747 | 0.567 | 0.401 | ||
| KernelSHAP | 0.662 | 0.275 | 0.194 | 0.765 | 0.570 | 0.398 | ||
| S1-SHAP (ours) | 0.666 | 0.299 | 0.195 | 0.732 | 0.586 | 0.419 | ||
| Random | 0.226 | 0.513 | 0.488 | 0.507 | 0.317 | 0.507 | ||
| Task | Primitive | Identifier | Question Instruction | Option Criteria / Target Semantics |
|---|---|---|---|---|
| e-SNLI | Choice ( ) | label | “Given that the premise is true, what is the relation of the hypothesis to the premise?” | entailment : The hypothesis is definitely true given the premise. neutral : The hypothesis might be true given the premise. contradiction : The hypothesis is definitely false given the premise. |
| Noul ( ) | noul | “Given that the premise is true, is the hypothesis definitely true?” | true : The premise entails the hypothesis. false : The premise does not entail the hypothesis. | |
| HANS | Choice ( ) | label | (Identical to e-SNLI; probabilities of neutral and contradiction are summed into non_entailment ). | (Identical to e-SNLI). |
| Noul ( ) | noul | “Given that the premise is true, is the hypothesis definitely true?” | true : The premise entails the hypothesis. false : The premise does not entail the hypothesis. | |
| Clinical Triage | Choice ( ) | label | “What emergency triage level does this patient require?” | immediate : Life-threatening condition requiring immediate resuscitation (ESI 1). urgent : Serious but stable condition that should be seen within an hour (ESI 2–3). non_urgent : Minor condition that can safely wait (ESI 4–5). |
| Noul ( ) | noul | “Does the patient require immediate care?” | true : The patient requires immediate emergency care. false : The patient does not require immediate emergency care. |