# Language Modeling on Pythia-160M Source: [Language Modeling on Pythia-160M]() Collection index: [llms.txt]() Data updated: 2026-04-28T07:00:07.703Z Evaluation of Language Modeling performance on the Pythia-160M dataset, reporting metrics such as Delta Perplexity (ΔPPL) and Log-Normalized Zero-shot Accuracy (L-N/ZA). A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Language Modeling | | Dataset (including any stated split) | Pythia-160M | | Metric | Delta Perplexity (ΔPPL) | | Unit | SCALAR | | Direction | Lower is better | | Results loaded | 16 of 16 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [Delta Perplexity (ΔPPL)](): Lower is better; unit: SCALAR. - [Log-Normalized Zero-shot Accuracy (L-N/ZA)](): Higher is better; unit: SCALAR. ## Recorded results | Method | Delta Perplexity (ΔPPL) | Result date | Report paper | | --- | --- | --- | --- | | Worst | 14.47 | Not recorded | 69f03f0755a3c32af89f9f94 | | Random | 14.5 | Not recorded | 69f03f0755a3c32af89f9f94 | | Zone-Aware Spectral Layer Pruning | 35.21 | Not recorded | 69f03f0755a3c32af89f9f94 | | Last-N | 43.14 | Not recorded | 69f03f0755a3c32af89f9f94 | | Zone-Aware Spectral Layer Pruning | 101.03 | Not recorded | 69f03f0755a3c32af89f9f94 | | Last-N | 111.94 | Not recorded | 69f03f0755a3c32af89f9f94 | | Random | 123.59 | Not recorded | 69f03f0755a3c32af89f9f94 | | Zone-Aware Spectral Layer Pruning | 195.23 | Not recorded | 69f03f0755a3c32af89f9f94 | | Last-N | 389.77 | Not recorded | 69f03f0755a3c32af89f9f94 | | Zone-Aware Spectral Layer Pruning | 501.59 | Not recorded | 69f03f0755a3c32af89f9f94 | | Random | 591.06 | Not recorded | 69f03f0755a3c32af89f9f94 | | Last-N | 633.19 | Not recorded | 69f03f0755a3c32af89f9f94 | | Random | 1224 | Not recorded | 69f03f0755a3c32af89f9f94 | | Worst | 34875 | Not recorded | 69f03f0755a3c32af89f9f94 | | Worst | 156415 | Not recorded | 69f03f0755a3c32af89f9f94 | | Worst | 5959385 | Not recorded | 69f03f0755a3c32af89f9f94 | ### Worst Delta Perplexity (ΔPPL): 14.47. Result ID: 69f05a1847f34ce924d5a4ec. [Reporting paper]() Recorded evaluation configuration: {"k":"1"}. This record does not assert independent reproduction. ### Random Delta Perplexity (ΔPPL): 14.5. Result ID: 69f05a1847f34ce924d5a4eb. [Reporting paper]() Recorded evaluation configuration: {"k":"1"}. This record does not assert independent reproduction. ### Zone-Aware Spectral Layer Pruning Delta Perplexity (ΔPPL): 35.21. Result ID: 69f05a1847f34ce924d5a4e9. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"k":"1"}. This record does not assert independent reproduction. ### Last-N Delta Perplexity (ΔPPL): 43.14. Result ID: 69f05a1847f34ce924d5a4ea. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"k":"1"}. This record does not assert independent reproduction. ### Zone-Aware Spectral Layer Pruning Delta Perplexity (ΔPPL): 101.03. Result ID: 69f05a1847f34ce924d5a4ed. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"k":"2"}. This record does not assert independent reproduction. ### Last-N Delta Perplexity (ΔPPL): 111.94. Result ID: 69f05a1847f34ce924d5a4ee. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"k":"2"}. This record does not assert independent reproduction. ### Random Delta Perplexity (ΔPPL): 123.59. Result ID: 69f05a1847f34ce924d5a4ef. [Reporting paper]() Recorded evaluation configuration: {"k":"2"}. This record does not assert independent reproduction. ### Zone-Aware Spectral Layer Pruning Delta Perplexity (ΔPPL): 195.23. Result ID: 69f05a1847f34ce924d5a4f1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"k":"3"}. This record does not assert independent reproduction. ### Last-N Delta Perplexity (ΔPPL): 389.77. Result ID: 69f05a1847f34ce924d5a4f2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"k":"3"}. This record does not assert independent reproduction. ### Zone-Aware Spectral Layer Pruning Delta Perplexity (ΔPPL): 501.59. Result ID: 69f05a1847f34ce924d5a4f5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"k":"4"}. This record does not assert independent reproduction. ### Random Delta Perplexity (ΔPPL): 591.06. Result ID: 69f05a1847f34ce924d5a4f3. [Reporting paper]() Recorded evaluation configuration: {"k":"3"}. This record does not assert independent reproduction. ### Last-N Delta Perplexity (ΔPPL): 633.19. Result ID: 69f05a1847f34ce924d5a4f6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"k":"4"}. This record does not assert independent reproduction. ### Random Delta Perplexity (ΔPPL): 1224. Result ID: 69f05a1847f34ce924d5a4f7. [Reporting paper]() Recorded evaluation configuration: {"k":"4"}. This record does not assert independent reproduction. ### Worst Delta Perplexity (ΔPPL): 34875. Result ID: 69f05a1847f34ce924d5a4f0. [Reporting paper]() Recorded evaluation configuration: {"k":"2"}. This record does not assert independent reproduction. ### Worst Delta Perplexity (ΔPPL): 156415. Result ID: 69f05a1847f34ce924d5a4f4. [Reporting paper]() Recorded evaluation configuration: {"k":"3"}. This record does not assert independent reproduction. ### Worst Delta Perplexity (ΔPPL): 5959385. Result ID: 69f05a1847f34ce924d5a4f8. [Reporting paper]() Recorded evaluation configuration: {"k":"4"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.