# Diverse Language Understanding on 62 downstream tasks Source: [Diverse Language Understanding on 62 downstream tasks]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:35:17.328Z Evaluation of model performance across 62 diverse downstream tasks focusing on general language understanding capabilities, typically measured by Average Accuracy. 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 | Diverse Language Understanding | | Dataset (including any stated split) | 62 downstream tasks | | Metric | Average Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 18 of 18 | | 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 - [Average Accuracy](): Higher is better; unit: PERCENT. ## Recorded results | Method | Average Accuracy | Result date | Report paper | | --- | --- | --- | --- | | Ours | 67.5 | Not recorded | 6992a4a19ebc499c97406ca8 | | AdaMerging | 66.9 | Not recorded | 6992a4a19ebc499c97406ca8 | | π-Tuning | 66.8 | Not recorded | 6992a4a19ebc499c97406ca8 | | π-Tuning | 66.8 | Not recorded | 6992a4a19ebc499c97406ca8 | | LoraHub | 66.3 | Not recorded | 6992a4a19ebc499c97406ca8 | | LoRA | 65.7 | Not recorded | 6992a4a19ebc499c97406ca8 | | AdaMerging | 65.2 | Not recorded | 6992a4a19ebc499c97406ca8 | | Ours | 65.2 | Not recorded | 6992a4a19ebc499c97406ca8 | | LoraHub | 56.3 | Not recorded | 6992a4a19ebc499c97406ca8 | | Simple avg. | 53.6 | Not recorded | 6992a4a19ebc499c97406ca8 | | TSV | 52.2 | Not recorded | 6992a4a19ebc499c97406ca8 | | Simple avg. | 49.6 | Not recorded | 6992a4a19ebc499c97406ca8 | | Simple avg. | 49.6 | Not recorded | 6992a4a19ebc499c97406ca8 | | Simple avg. | 49.5 | Not recorded | 6992a4a19ebc499c97406ca8 | | TSV | 48.8 | Not recorded | 6992a4a19ebc499c97406ca8 | | TIES | 46.7 | Not recorded | 6992a4a19ebc499c97406ca8 | | Prompting | 46.7 | Not recorded | 6992a4a19ebc499c97406ca8 | | TIES | 46.6 | Not recorded | 6992a4a19ebc499c97406ca8 | ### Ours Average Accuracy: 67.5. Result ID: 699a4cfc5698665ec4d48204. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LoRA pool size":"30","Target-task LoRA included":"true"}. This record does not assert independent reproduction. ### AdaMerging Average Accuracy: 66.9. Result ID: 699a4cfc5698665ec4d481fe. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LoRA pool size":"30","Target-task LoRA included":"true"}. This record does not assert independent reproduction. ### π-Tuning Average Accuracy: 66.8. Result ID: 699a4cfc5698665ec4d481ff. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LoRA pool size":"20","Target-task LoRA included":"false"}. This record does not assert independent reproduction. ### π-Tuning Average Accuracy: 66.8. Result ID: 699a4cfc5698665ec4d48200. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LoRA pool size":"20","Target-task LoRA included":"true"}. This record does not assert independent reproduction. ### LoraHub Average Accuracy: 66.3. Result ID: 699a4cfc5698665ec4d48202. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LoRA pool size":"30","Target-task LoRA included":"true"}. This record does not assert independent reproduction. ### LoRA Average Accuracy: 65.7. Result ID: 699a4cfc5698665ec4d48206. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Training":"Trained on target-task data"}. This record does not assert independent reproduction. ### AdaMerging Average Accuracy: 65.2. Result ID: 699a4cfc5698665ec4d481fd. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LoRA pool size":"30","Target-task LoRA included":"false"}. This record does not assert independent reproduction. ### Ours Average Accuracy: 65.2. Result ID: 699a4cfc5698665ec4d48203. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LoRA pool size":"30","Target-task LoRA included":"false"}. This record does not assert independent reproduction. ### LoraHub Average Accuracy: 56.3. Result ID: 699a4cfc5698665ec4d48201. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LoRA pool size":"30","Target-task LoRA included":"false"}. This record does not assert independent reproduction. ### Simple avg. Average Accuracy: 53.6. Result ID: 699a4cfc5698665ec4d481fa. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"LoRA pool size":"30","Target-task LoRA included":"true"}. This record does not assert independent reproduction. ### TSV Average Accuracy: 52.2. Result ID: 699a4cfc5698665ec4d481f6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LoRA pool size":"30","Target-task LoRA included":"true"}. This record does not assert independent reproduction. ### Simple avg. Average Accuracy: 49.6. Result ID: 699a4cfc5698665ec4d481f9. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"LoRA pool size":"30","Target-task LoRA included":"false"}. This record does not assert independent reproduction. ### Simple avg. Average Accuracy: 49.6. Result ID: 699a4cfc5698665ec4d481fc. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"LoRA pool size":"all","Target-task LoRA included":"true"}. This record does not assert independent reproduction. ### Simple avg. Average Accuracy: 49.5. Result ID: 699a4cfc5698665ec4d481fb. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"LoRA pool size":"all","Target-task LoRA included":"false"}. This record does not assert independent reproduction. ### TSV Average Accuracy: 48.8. Result ID: 699a4cfc5698665ec4d481f5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LoRA pool size":"30","Target-task LoRA included":"false"}. This record does not assert independent reproduction. ### TIES Average Accuracy: 46.7. Result ID: 699a4cfc5698665ec4d481f7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LoRA pool size":"30","Target-task LoRA included":"false"}. This record does not assert independent reproduction. ### Prompting Average Accuracy: 46.7. Result ID: 699a4cfc5698665ec4d48205. [Reporting paper]() Recorded evaluation configuration: {"protocol":"zero-shot"}. This record does not assert independent reproduction. ### TIES Average Accuracy: 46.6. Result ID: 699a4cfc5698665ec4d481f8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LoRA pool size":"30","Target-task LoRA included":"true"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.