# Temporal Reasoning on GDELT (test) Source: [Temporal Reasoning on GDELT (test)]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:58:09.955Z Evaluation of Temporal Reasoning capabilities on the test split of the GDELT dataset, reporting performance metrics across Positive, Negative, and Neutral outcomes. 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 | Temporal Reasoning | | Dataset (including any stated split) | GDELT (test) | | Metric | Positive Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 17 of 17 | | 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 - [Positive Accuracy](): Higher is better; unit: PERCENT. - [Negative Accuracy](): Higher is better; unit: PERCENT. - [Neutral Accuracy](): Higher is better; unit: PERCENT. - [Overall Accuracy](): Higher is better; unit: PERCENT. ## Recorded results | Method | Positive Accuracy | Result date | Report paper | | --- | --- | --- | --- | | GETER | 63.77 | Not recorded | 68fa78a61e209131cea799af | | Llama3-8B-Instruct | 62.86 | Not recorded | 68fa78a61e209131cea799af | | GETER | 62.62 | Not recorded | 68fa78a61e209131cea799af | | Llama3-8B-Instruct | 61.94 | Not recorded | 68fa78a61e209131cea799af | | GETER | 61.29 | Not recorded | 68fa78a61e209131cea799af | | Mistral-7B-Instruct | 60.56 | Not recorded | 68fa78a61e209131cea799af | | Llama3-8B-Instruct | 53.48 | Not recorded | 68fa78a61e209131cea799af | | GPT-4o | 42.74 | Not recorded | 68fa78a61e209131cea799af | | Qwen2.5-7B-Instruct | 31.28 | Not recorded | 68fa78a61e209131cea799af | | Qwen2.5-7B-Instruct | 22.39 | Not recorded | 68fa78a61e209131cea799af | | Qwen2.5-7B-Instruct | 22.22 | Not recorded | 68fa78a61e209131cea799af | | Mistral-7B-Instruct | 19.45 | Not recorded | 68fa78a61e209131cea799af | | GPT-4o | 19.08 | Not recorded | 68fa78a61e209131cea799af | | Qwen2.5-7B-Instruct | 11.27 | Not recorded | 68fa78a61e209131cea799af | | Mistral-7B-Instruct | 5.52 | Not recorded | 68fa78a61e209131cea799af | | Llama3-8B-Instruct | 1.95 | Not recorded | 68fa78a61e209131cea799af | | Mistral-7B-Instruct | 0.99 | Not recorded | 68fa78a61e209131cea799af | ### GETER Positive Accuracy: 63.77. Result ID: 699598913badfaffe693d09b. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B-Instruct"}. This record does not assert independent reproduction. ### Llama3-8B-Instruct Positive Accuracy: 62.86. Result ID: 699598913badfaffe693d095. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Protocol":"LoRA","Reasoning Chains":true}. This record does not assert independent reproduction. ### GETER Positive Accuracy: 62.62. Result ID: 699598913badfaffe693d096. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Llama3-8B-Instruct"}. This record does not assert independent reproduction. ### Llama3-8B-Instruct Positive Accuracy: 61.94. Result ID: 699598913badfaffe693d094. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Protocol":"LoRA","Reasoning Chains":false}. This record does not assert independent reproduction. ### GETER Positive Accuracy: 61.29. Result ID: 699598913badfaffe693d0a0. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Mistral-7B-Instruct"}. This record does not assert independent reproduction. ### Mistral-7B-Instruct Positive Accuracy: 60.56. Result ID: 699598913badfaffe693d09f. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Protocol":"LoRA","Reasoning Chains":true}. This record does not assert independent reproduction. ### Llama3-8B-Instruct Positive Accuracy: 53.48. Result ID: 699598913badfaffe693d093. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Protocol":"zero-shot","Reasoning Chains":true}. This record does not assert independent reproduction. ### GPT-4o Positive Accuracy: 42.74. Result ID: 699598913badfaffe693d091. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Protocol":"zero-shot","Reasoning Chains":true}. This record does not assert independent reproduction. ### Qwen2.5-7B-Instruct Positive Accuracy: 31.28. Result ID: 699598913badfaffe693d099. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Protocol":"LoRA","Reasoning Chains":false}. This record does not assert independent reproduction. ### Qwen2.5-7B-Instruct Positive Accuracy: 22.39. Result ID: 699598913badfaffe693d09a. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Protocol":"LoRA","Reasoning Chains":true}. This record does not assert independent reproduction. ### Qwen2.5-7B-Instruct Positive Accuracy: 22.22. Result ID: 699598913badfaffe693d098. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Protocol":"zero-shot","Reasoning Chains":true}. This record does not assert independent reproduction. ### Mistral-7B-Instruct Positive Accuracy: 19.45. Result ID: 699598913badfaffe693d09e. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Protocol":"LoRA","Reasoning Chains":false}. This record does not assert independent reproduction. ### GPT-4o Positive Accuracy: 19.08. Result ID: 699598913badfaffe693d090. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Protocol":"zero-shot","Reasoning Chains":false}. This record does not assert independent reproduction. ### Qwen2.5-7B-Instruct Positive Accuracy: 11.27. Result ID: 699598913badfaffe693d097. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Protocol":"zero-shot","Reasoning Chains":false}. This record does not assert independent reproduction. ### Mistral-7B-Instruct Positive Accuracy: 5.52. Result ID: 699598913badfaffe693d09c. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Protocol":"zero-shot","Reasoning Chains":false}. This record does not assert independent reproduction. ### Llama3-8B-Instruct Positive Accuracy: 1.95. Result ID: 699598913badfaffe693d092. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Protocol":"zero-shot","Reasoning Chains":false}. This record does not assert independent reproduction. ### Mistral-7B-Instruct Positive Accuracy: 0.99. Result ID: 699598913badfaffe693d09d. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Protocol":"zero-shot","Reasoning Chains":true}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.