Meaning-preservation results
Semantic preservation, without using detector scores. All values come from our shared evaluation.
Measurement coverage5,252 / 5,313
Current scores
Scores are provisional while 58 detector checks remain unresolved. Rankings update as measurements arrive. Provider exclusions: 3. Excluded checks have no numeric score. Readability and meaning: 759/759 recorded.
Inspect coverage and recorded evidence →| Entry | Raw cosine | Rescaled meaning | Readability | Evidence |
|---|---|---|---|---|
Open Humanizer 4BVerified by HumanizeBenchWe ran this model ourselves on the benchmark source examples and checked its execution records. Verification confirms the tested model’s provenance; it does not guarantee performance or imply that every detector returned a result.akhaliq/open-humanizer-4b | 0.9916 | 96.6 | 85.1 | Read outputs → |
GoHumanizeVerified by HumanizeBenchWe ran this model ourselves on the benchmark source examples and checked its execution records. Verification confirms the tested model’s provenance; it does not guarantee performance or imply that every detector returned a result.gohumanize/gohumanize-open-humanizer | 0.9801 | 92.0 | 85.1 | Read outputs → |
Smaug HumanizerVerified by HumanizeBenchWe ran this model ourselves on the benchmark source examples and checked its execution records. Verification confirms the tested model’s provenance; it does not guarantee performance or imply that every detector returned a result.abacusai/Smaug-Humanizer | 0.9414 | 76.5 | 81.3 | Read outputs → |
HumanizerAI · GGUFVerified by HumanizeBenchWe ran this model ourselves on the benchmark source examples and checked its execution records. Verification confirms the tested model’s provenance; it does not guarantee performance or imply that every detector returned a result.mradermacher/humanizerAI-GGUF | 0.9382 | 75.9 | 82.6 | Read outputs → |
Humanizer 1B · OptiQ 4-bitVerified by HumanizeBenchWe ran this model ourselves on the benchmark source examples and checked its execution records. Verification confirms the tested model’s provenance; it does not guarantee performance or imply that every detector returned a result.mlx-community/humanizer-1B-OptiQ-4bit | 0.9242 | 77.1 | 71.2 | Read outputs → |
Humanizer 12B · GGUFVerified by HumanizeBenchWe ran this model ourselves on the benchmark source examples and checked its execution records. Verification confirms the tested model’s provenance; it does not guarantee performance or imply that every detector returned a result.jialinyyzz/humanizer-GGUF | 0.9241 | 69.6 | 80.0 | Read outputs → |
Humanizer 12BVerified by HumanizeBenchWe ran this model ourselves on the benchmark source examples and checked its execution records. Verification confirms the tested model’s provenance; it does not guarantee performance or imply that every detector returned a result.jialinyyzz/humanizer | 0.9241 | 69.6 | 80.0 | Read outputs → |
Humanizer 12B · MLX 8-bit (cowWhySo)Verified by HumanizeBenchWe ran this model ourselves on the benchmark source examples and checked its execution records. Verification confirms the tested model’s provenance; it does not guarantee performance or imply that every detector returned a result.cowWhySo/humanizer-12b-mlx-8bit | 0.9163 | 66.5 | 79.0 | Read outputs → |
Humanizer · CMFVerified by HumanizeBenchWe ran this model ourselves on the benchmark source examples and checked its execution records. Verification confirms the tested model’s provenance; it does not guarantee performance or imply that every detector returned a result.infosave/humanizer-cmf | 0.9147 | 65.9 | 78.0 | Read outputs → |
Humanizer 12B · MLX 8-bit (npario)Verified by HumanizeBenchWe ran this model ourselves on the benchmark source examples and checked its execution records. Verification confirms the tested model’s provenance; it does not guarantee performance or imply that every detector returned a result.npario/humanizer-12b-mlx-8bit | 0.9147 | 65.9 | 78.8 | Read outputs → |
Humanizer 12B · OptiQ 4-bitVerified by HumanizeBenchWe ran this model ourselves on the benchmark source examples and checked its execution records. Verification confirms the tested model’s provenance; it does not guarantee performance or imply that every detector returned a result.mlx-community/humanizer-12b-OptiQ-4bit | 0.9022 | 60.9 | 77.1 | Read outputs → |
| 0.8987 | 60.2 | 82.2 | Read outputs → | |
| 0.8856 | 55.6 | 79.6 | Read outputs → | |
| 0.8755 | 50.2 | 77.6 | Read outputs → | |
| 0.8661 | 46.8 | 79.8 | Read outputs → | |
| 0.8644 | 46.4 | 79.9 | Read outputs → | |
| 0.8522 | 41.3 | 76.9 | Read outputs → | |
| 0.8522 | 43.2 | 74.6 | Read outputs → | |
| 0.8518 | 41.4 | 71.7 | Read outputs → | |
| 0.8328 | 35.7 | 73.4 | Read outputs → | |
| 0.8061 | 25.4 | 75.8 | Read outputs → | |
Qwen 2.5 14B HumanizerVerified by HumanizeBenchWe ran this model ourselves on the benchmark source examples and checked its execution records. Verification confirms the tested model’s provenance; it does not guarantee performance or imply that every detector returned a result.arunsingh80475/qwen2.5-14b-humanizer-v4-lora | 0.7892 | 21.1 | 78.1 | Read outputs → |
Humanizer 3BVerified by HumanizeBenchWe ran this model ourselves on the benchmark source examples and checked its execution records. Verification confirms the tested model’s provenance; it does not guarantee performance or imply that every detector returned a result.mondk/Humanizer-3B-Safetensors | 0.7580 | 12.3 | 72.1 | Read outputs → |