# Implicit Sentiment Analysis on SemEval Laptop ISA 2014 Source: [Implicit Sentiment Analysis on SemEval Laptop ISA 2014]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:43:16.204Z Evaluation of Implicit Sentiment Analysis performance on the SemEval 2014 Task 4 Laptop dataset, measured using F1 score. 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 | Implicit Sentiment Analysis | | Dataset (including any stated split) | SemEval Laptop ISA 2014 | | Metric | F1 Score | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 13 of 13 | | 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 - [F1 Score](): Higher is better; unit: PERCENT. ## Recorded results | Method | F1 Score | Result date | Report paper | | --- | --- | --- | --- | | THOR | 82.43 | Not recorded | 68fa6fbc1e209131cea471d1 | | THOR | 81.69 | Not recorded | 68fa6fbc1e209131cea471d1 | | THOR | 80.27 | Not recorded | 68fa6fbc1e209131cea471d1 | | Flan-T5+Prompt | 78.91 | Not recorded | 68fa6fbc1e209131cea471d1 | | BERT+ISAIV | 78.29 | Not recorded | 68fa6fbc1e209131cea471d1 | | BERTAsp+SCAPT | 77.59 | Not recorded | 68fa6fbc1e209131cea471d1 | | Flan-T5+Prompt | 76.4 | Not recorded | 68fa6fbc1e209131cea471d1 | | BERTAsp+CEPT | 75.86 | Not recorded | 68fa6fbc1e209131cea471d1 | | BERT+Prompt | 75.24 | Not recorded | 68fa6fbc1e209131cea471d1 | | BERT+RGAT | 72.99 | Not recorded | 68fa6fbc1e209131cea471d1 | | BERT+ADA | 70.11 | Not recorded | 68fa6fbc1e209131cea471d1 | | BERT+SPC | 69.54 | Not recorded | 68fa6fbc1e209131cea471d1 | | THOR | 67.63 | Not recorded | 68fa6fbc1e209131cea471d1 | ### THOR F1 Score: 82.43. Result ID: 6990e8fa3a6c31484904377b. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Parameters":"11B","Backbone":"Flan-T5","Evaluation Protocol":"Supervised fine-tuning"}. This record does not assert independent reproduction. ### THOR F1 Score: 81.69. Result ID: 6990e8fa3a6c31484904377d. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Parameters":"11B","Backbone":"Flan-T5","Evaluation Protocol":"Supervised fine-tuning","Method Variation":"w/o Reason-revising"}. This record does not assert independent reproduction. ### THOR F1 Score: 80.27. Result ID: 6990e8fa3a6c31484904377c. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Parameters":"11B","Backbone":"Flan-T5","Evaluation Protocol":"Supervised fine-tuning","Method Variation":"w/o Self-consistency"}. This record does not assert independent reproduction. ### Flan-T5+Prompt F1 Score: 78.91. Result ID: 6990e8fa3a6c314849043779. [Reporting paper]() Recorded evaluation configuration: {"Model Parameters":"11B","Backbone":"Flan-T5","Evaluation Protocol":"Supervised fine-tuning"}. This record does not assert independent reproduction. ### BERT+ISAIV F1 Score: 78.29. Result ID: 6990e8fa3a6c314849043775. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Parameters":"110M","Backbone":"BERT","Evaluation Protocol":"Supervised fine-tuning"}. This record does not assert independent reproduction. ### BERTAsp+SCAPT F1 Score: 77.59. Result ID: 6990e8fa3a6c314849043776. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Parameters":"110M","Backbone":"BERT","Evaluation Protocol":"Supervised fine-tuning"}. This record does not assert independent reproduction. ### Flan-T5+Prompt F1 Score: 76.4. Result ID: 6990e8fa3a6c314849043778. [Reporting paper]() Recorded evaluation configuration: {"Model Parameters":"250M","Backbone":"Flan-T5","Evaluation Protocol":"Supervised fine-tuning"}. This record does not assert independent reproduction. ### BERTAsp+CEPT F1 Score: 75.86. Result ID: 6990e8fa3a6c314849043774. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Parameters":"110M","Backbone":"BERT","Evaluation Protocol":"Supervised fine-tuning"}. This record does not assert independent reproduction. ### BERT+Prompt F1 Score: 75.24. Result ID: 6990e8fa3a6c314849043777. [Reporting paper]() Recorded evaluation configuration: {"Model Parameters":"110M","Backbone":"BERT","Evaluation Protocol":"Supervised fine-tuning"}. This record does not assert independent reproduction. ### BERT+RGAT F1 Score: 72.99. Result ID: 6990e8fa3a6c314849043773. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Parameters":"110M","Backbone":"BERT","Evaluation Protocol":"Supervised fine-tuning"}. This record does not assert independent reproduction. ### BERT+ADA F1 Score: 70.11. Result ID: 6990e8fa3a6c314849043772. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Parameters":"110M","Backbone":"BERT","Evaluation Protocol":"Supervised fine-tuning"}. This record does not assert independent reproduction. ### BERT+SPC F1 Score: 69.54. Result ID: 6990e8fa3a6c314849043771. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Parameters":"110M","Backbone":"BERT","Evaluation Protocol":"Supervised fine-tuning"}. This record does not assert independent reproduction. ### THOR F1 Score: 67.63. Result ID: 6990e8fa3a6c31484904377a. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Parameters":"250M","Backbone":"Flan-T5","Evaluation Protocol":"Supervised fine-tuning"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.