PHASE 1 RESEARCH ARTIFACT — detects model confidence, not factual correctness

This model was trained on LLM hidden-state activations to produce an energy score that correlates with the model's output confidence (hallucination likelihood). It cannot verify whether a model's answer is factually correct — it can only signal how uncertain the model appears token-by-token.

This limitation was confirmed in Exp 184/203: the energy scores reflect model confidence, not answer correctness. Do not use these scores as a correctness verifier.

For production use, install the full Carnot pipeline:

pip install carnot-ebm

The production pipeline includes FormalClaimVerifier (solver-routed formal claim verification), PBT code verification, and the Carnot MCP server. See Carnot on GitHub for documentation.

Exp 316 Full-Scale Benchmark Results (2026-04-14)

• energy-based-model
• hallucination-detection
• jax
• carnot license: apache-2.0

⚠️ PHASE 1 RESEARCH ARTIFACT

This model detects output confidence (hallucination likelihood signals from LLM hidden-state activations), not correctness. It cannot verify whether a model's answer is right — it can only signal how uncertain the model appears token-by-token.

For production use, install the full Carnot pipeline which includes FormalClaimVerifier (solver-routed formal claim verification), PBT code verification (property-based testing on 164-problem HumanEval), process integrity detection (right-for-wrong-reasons), and the Carnot MCP server:

pip install carnot-ebm

See Carnot on GitHub for documentation and the full production API.

Note: These are Phase 1 research artifacts — per-token activation EBMs that detect hallucination confidence signals from LLM hidden states. For production use of the full Carnot EBM framework (constraint verification, guided decoding, energy-based repair), see:

pip install carnot-ebm

Source and documentation: https://github.com/Carnot-EBM/carnot-ebm

Important: Research Artifact, Not a Production Detector

This model achieves 83.4% on held-out TruthfulQA test sets, but in practical deployment (8 real questions), activation-based EBMs agreed with ground truth only 50% of the time. The EBM detects model confidence, not correctness — confident hallucinations get low energy (look fine) while correct-but-hedging answers get flagged.

This model is a research artifact documenting activation-space structure. It is NOT a reliable hallucination detector for production use.

For practical verification, use structural constraints (test execution, SAT solving) rather than activation analysis. See the Carnot technical report for 41 experiments and 14 principles learned.

per-token-ebm-qwen3-06b

Key Stats

Usage

Training

• Loss: Noise Contrastive Estimation (NCE)
• Epochs: 300, lr=0.005
• Data: 26,800 per-token activations from Qwen/Qwen3-0.6B
• Labels: correct answer tokens = low energy (data), wrong answer tokens = high energy (noise)

Limitations

• Only works with activations from Qwen/Qwen3-0.6B — different models have different representation spaces
• 83.4% accuracy — use as one signal among many, not as sole verification
• Trained on QA/TruthfulQA — may not generalize to all domains

10 Principles from Carnot Research

1. Simpler is better in small-data regimes
2. Token-level features > sequence-level
3. The model's own logprobs are the best energy
4. Overfitting is the main enemy
5. Extract features from generated tokens, not prompts
6. Different energy signals dominate in different domains
7. Statistical difference ≠ causal influence
8. Instruction tuning compresses the hallucination signal
9. Adversarial questions defeat post-hoc detection
10. Chain-of-thought compresses the hallucination signal

What's Proven to Work (2026)

Production Use