The Truncation Blind Spot: How Decoding Strategies Systematically Exclude Human-Like Token Choices
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866910059301175296 |
|---|---|
| author | Arias, Esteban Garces Sapargali, Nurzhan Heumann, Christian Aßenmacher, Matthias |
| author_facet | Arias, Esteban Garces Sapargali, Nurzhan Heumann, Christian Aßenmacher, Matthias |
| contents | Standard decoding strategies for text generation, including top-k, nucleus sampling, and contrastive search, select tokens based on likelihood, restricting selection to high-probability regions. Human language production operates differently: tokens are chosen for communicative appropriateness rather than statistical frequency. This mismatch creates a truncation blind spot: contextually appropriate but statistically rare tokens remain accessible to humans yet unreachable by likelihood-based decoding. We hypothesize this contributes to the detectability of machine-generated text. Analyzing over 1.8 million texts across eight language models, five decoding strategies, and 53 hyperparameter configurations, we find that 8-18% of human-selected tokens fall outside typical truncation boundaries. Simple classifiers trained on predictability and lexical diversity achieve remarkable detection rates. Crucially, neither model scale nor architecture correlates strongly with detectability; truncation parameters account for most variance. Configurations achieving low detectability often produce incoherent text, indicating that evading detection and producing natural text are distinct objectives. These findings suggest detectability is enhanced by likelihood-based token selection, not merely a matter of model capability. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_18482 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | The Truncation Blind Spot: How Decoding Strategies Systematically Exclude Human-Like Token Choices Arias, Esteban Garces Sapargali, Nurzhan Heumann, Christian Aßenmacher, Matthias Computation and Language Machine Learning Standard decoding strategies for text generation, including top-k, nucleus sampling, and contrastive search, select tokens based on likelihood, restricting selection to high-probability regions. Human language production operates differently: tokens are chosen for communicative appropriateness rather than statistical frequency. This mismatch creates a truncation blind spot: contextually appropriate but statistically rare tokens remain accessible to humans yet unreachable by likelihood-based decoding. We hypothesize this contributes to the detectability of machine-generated text. Analyzing over 1.8 million texts across eight language models, five decoding strategies, and 53 hyperparameter configurations, we find that 8-18% of human-selected tokens fall outside typical truncation boundaries. Simple classifiers trained on predictability and lexical diversity achieve remarkable detection rates. Crucially, neither model scale nor architecture correlates strongly with detectability; truncation parameters account for most variance. Configurations achieving low detectability often produce incoherent text, indicating that evading detection and producing natural text are distinct objectives. These findings suggest detectability is enhanced by likelihood-based token selection, not merely a matter of model capability. |
| title | The Truncation Blind Spot: How Decoding Strategies Systematically Exclude Human-Like Token Choices |
| topic | Computation and Language Machine Learning |
| url | https://arxiv.org/abs/2603.18482 |