# Downstream Task Accuracy via Paradigm Routing on Downstream Tasks (test) Source: [Downstream Task Accuracy via Paradigm Routing on Downstream Tasks (test)]() Collection index: [llms.txt]() Data updated: 2026-04-10T00:18:13.861Z Evaluation of model accuracy on general downstream tasks utilizing the Paradigm Routing mechanism, reported on the test split. 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 | Downstream Task Accuracy via Paradigm Routing | | Dataset (including any stated split) | Downstream Tasks (test) | | Metric | Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 45 of 45 | | 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 - [Accuracy](): Higher is better; unit: PERCENT. - [Average Score](): Higher is better; unit: PERCENT. ## Recorded results | Method | Accuracy | Result date | Report paper | | --- | --- | --- | --- | | Oracle | 73.4 | Not recorded | 69d732b67d0172140ec76cc2 | | Oracle | 72.9 | Not recorded | 69d732b67d0172140ec76cc2 | | Oracle | 72.5 | Not recorded | 69d732b67d0172140ec76cc2 | | Embedding (LR) | 64.2 | Not recorded | 69d732b67d0172140ec76cc2 | | Combined (LR) | 64.2 | Not recorded | 69d732b67d0172140ec76cc2 | | Best-single | 62.4 | Not recorded | 69d732b67d0172140ec76cc2 | | Handcrafted (MLP) | 62.4 | Not recorded | 69d732b67d0172140ec76cc2 | | Combined (MLP) | 62.4 | Not recorded | 69d732b67d0172140ec76cc2 | | Combined (MLP) | 62.2 | Not recorded | 69d732b67d0172140ec76cc2 | | Embedding (LR) | 61.9 | Not recorded | 69d732b67d0172140ec76cc2 | | Combined (LR) | 61 | Not recorded | 69d732b67d0172140ec76cc2 | | Direct | 60.3 | Not recorded | 69d732b67d0172140ec76cc2 | | Self-route | 60.3 | Not recorded | 69d732b67d0172140ec76cc2 | | Handcrafted (LR) | 59.4 | Not recorded | 69d732b67d0172140ec76cc2 | | Handcrafted (LR) | 58.5 | Not recorded | 69d732b67d0172140ec76cc2 | | Self-route | 56.8 | Not recorded | 69d732b67d0172140ec76cc2 | | Oracle | 56.8 | Not recorded | 69d732b67d0172140ec76cc2 | | Direct | 55.5 | Not recorded | 69d732b67d0172140ec76cc2 | | Best-single | 55.5 | Not recorded | 69d732b67d0172140ec76cc2 | | Handcrafted (MLP) | 55.5 | Not recorded | 69d732b67d0172140ec76cc2 | | Combined (LR) | 54.6 | Not recorded | 69d732b67d0172140ec76cc2 | | Handcrafted (MLP) | 53.3 | Not recorded | 69d732b67d0172140ec76cc2 | | Embedding (LR) | 53.3 | Not recorded | 69d732b67d0172140ec76cc2 | | Combined (MLP) | 53.3 | Not recorded | 69d732b67d0172140ec76cc2 | | Handcrafted (LR) | 51.5 | Not recorded | 69d732b67d0172140ec76cc2 | | Best-single | 50.7 | Not recorded | 69d732b67d0172140ec76cc2 | | Direct | 49.8 | Not recorded | 69d732b67d0172140ec76cc2 | | Self-route | 42.4 | Not recorded | 69d732b67d0172140ec76cc2 | | Handcrafted (MLP) | 37.1 | Not recorded | 69d732b67d0172140ec76cc2 | | Handcrafted (LR) | 34.9 | Not recorded | 69d732b67d0172140ec76cc2 | | Combined (MLP) | 33.6 | Not recorded | 69d732b67d0172140ec76cc2 | | Best-single | 32.8 | Not recorded | 69d732b67d0172140ec76cc2 | | Combined (LR) | 32.8 | Not recorded | 69d732b67d0172140ec76cc2 | | Embedding (LR) | 32.3 | Not recorded | 69d732b67d0172140ec76cc2 | | Self-route | 27.5 | Not recorded | 69d732b67d0172140ec76cc2 | | Direct | 24.9 | Not recorded | 69d732b67d0172140ec76cc2 | ### Oracle Accuracy: 73.4. Result ID: 69d8412347f34ce924c04523. [Reporting paper]() Recorded evaluation configuration: {"Router Category":"Oracle","Model":"Gemini"}. This record does not assert independent reproduction. ### Oracle Accuracy: 72.9. Result ID: 69d8412347f34ce924c04522. [Reporting paper]() Recorded evaluation configuration: {"Router Category":"Oracle","Model":"GPT-5"}. This record does not assert independent reproduction. ### Oracle Accuracy: 72.5. Result ID: 69d8412347f34ce924c04524. [Reporting paper]() Recorded evaluation configuration: {"Router Category":"Oracle","Model":"Qwen3-Max"}. This record does not assert independent reproduction. ### Embedding (LR) Accuracy: 64.2. Result ID: 69d8412347f34ce924c0450e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Router Category":"Emb","Model":"GPT-5"}. This record does not assert independent reproduction. ### Combined (LR) Accuracy: 64.2. Result ID: 69d8412347f34ce924c04513. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Router Category":"Emb","Model":"GPT-5"}. This record does not assert independent reproduction. ### Best-single Accuracy: 62.4. Result ID: 69d8412347f34ce924c044ff. [Reporting paper]() Recorded evaluation configuration: {"Router Category":"Base","Model":"GPT-5"}. This record does not assert independent reproduction. ### Handcrafted (MLP) Accuracy: 62.4. Result ID: 69d8412347f34ce924c04509. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Router Category":"HC","Model":"GPT-5"}. This record does not assert independent reproduction. ### Combined (MLP) Accuracy: 62.4. Result ID: 69d8412347f34ce924c04518. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Router Category":"Emb","Model":"GPT-5"}. This record does not assert independent reproduction. ### Combined (MLP) Accuracy: 62.2. Result ID: 69d8412347f34ce924c04519. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Router Category":"Emb","Model":"Gemini"}. This record does not assert independent reproduction. ### Embedding (LR) Accuracy: 61.9. Result ID: 69d8412347f34ce924c0450f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Router Category":"Emb","Model":"Gemini"}. This record does not assert independent reproduction. ### Combined (LR) Accuracy: 61. Result ID: 69d8412347f34ce924c04514. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Router Category":"Emb","Model":"Gemini"}. This record does not assert independent reproduction. ### Direct Accuracy: 60.3. Result ID: 69d8412347f34ce924c044fa. [Reporting paper]() Recorded evaluation configuration: {"Router Category":"Base","Model":"GPT-5"}. This record does not assert independent reproduction. ### Self-route Accuracy: 60.3. Result ID: 69d8412347f34ce924c0451d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Router Category":"ZS","Model":"GPT-5"}. This record does not assert independent reproduction. ### Handcrafted (LR) Accuracy: 59.4. Result ID: 69d8412347f34ce924c04504. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Router Category":"HC","Model":"GPT-5"}. This record does not assert independent reproduction. ### Handcrafted (LR) Accuracy: 58.5. Result ID: 69d8412347f34ce924c04505. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Router Category":"HC","Model":"Gemini"}. This record does not assert independent reproduction. ### Self-route Accuracy: 56.8. Result ID: 69d8412347f34ce924c0451e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Router Category":"ZS","Model":"Gemini"}. This record does not assert independent reproduction. ### Oracle Accuracy: 56.8. Result ID: 69d8412347f34ce924c04525. [Reporting paper]() Recorded evaluation configuration: {"Router Category":"Oracle","Model":"Qwen3-30B"}. This record does not assert independent reproduction. ### Direct Accuracy: 55.5. Result ID: 69d8412347f34ce924c044fb. [Reporting paper]() Recorded evaluation configuration: {"Router Category":"Base","Model":"Gemini"}. This record does not assert independent reproduction. ### Best-single Accuracy: 55.5. Result ID: 69d8412347f34ce924c04500. [Reporting paper]() Recorded evaluation configuration: {"Router Category":"Base","Model":"Gemini"}. This record does not assert independent reproduction. ### Handcrafted (MLP) Accuracy: 55.5. Result ID: 69d8412347f34ce924c0450a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Router Category":"HC","Model":"Gemini"}. This record does not assert independent reproduction. ### Combined (LR) Accuracy: 54.6. Result ID: 69d8412347f34ce924c04515. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Router Category":"Emb","Model":"Qwen3-Max"}. This record does not assert independent reproduction. ### Handcrafted (MLP) Accuracy: 53.3. Result ID: 69d8412347f34ce924c0450b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Router Category":"HC","Model":"Qwen3-Max"}. This record does not assert independent reproduction. ### Embedding (LR) Accuracy: 53.3. Result ID: 69d8412347f34ce924c04510. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Router Category":"Emb","Model":"Qwen3-Max"}. This record does not assert independent reproduction. ### Combined (MLP) Accuracy: 53.3. Result ID: 69d8412347f34ce924c0451a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Router Category":"Emb","Model":"Qwen3-Max"}. This record does not assert independent reproduction. ### Handcrafted (LR) Accuracy: 51.5. Result ID: 69d8412347f34ce924c04506. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Router Category":"HC","Model":"Qwen3-Max"}. This record does not assert independent reproduction. ### Best-single Accuracy: 50.7. Result ID: 69d8412347f34ce924c04501. [Reporting paper]() Recorded evaluation configuration: {"Router Category":"Base","Model":"Qwen3-Max"}. This record does not assert independent reproduction. ### Direct Accuracy: 49.8. Result ID: 69d8412347f34ce924c044fc. [Reporting paper]() Recorded evaluation configuration: {"Router Category":"Base","Model":"Qwen3-Max"}. This record does not assert independent reproduction. ### Self-route Accuracy: 42.4. Result ID: 69d8412347f34ce924c0451f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Router Category":"ZS","Model":"Qwen3-Max"}. This record does not assert independent reproduction. ### Handcrafted (MLP) Accuracy: 37.1. Result ID: 69d8412347f34ce924c0450c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Router Category":"HC","Model":"Qwen3-30B"}. This record does not assert independent reproduction. ### Handcrafted (LR) Accuracy: 34.9. Result ID: 69d8412347f34ce924c04507. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Router Category":"HC","Model":"Qwen3-30B"}. This record does not assert independent reproduction. ### Combined (MLP) Accuracy: 33.6. Result ID: 69d8412347f34ce924c0451b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Router Category":"Emb","Model":"Qwen3-30B"}. This record does not assert independent reproduction. ### Best-single Accuracy: 32.8. Result ID: 69d8412347f34ce924c04502. [Reporting paper]() Recorded evaluation configuration: {"Router Category":"Base","Model":"Qwen3-30B"}. This record does not assert independent reproduction. ### Combined (LR) Accuracy: 32.8. Result ID: 69d8412347f34ce924c04516. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Router Category":"Emb","Model":"Qwen3-30B"}. This record does not assert independent reproduction. ### Embedding (LR) Accuracy: 32.3. Result ID: 69d8412347f34ce924c04511. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Router Category":"Emb","Model":"Qwen3-30B"}. This record does not assert independent reproduction. ### Self-route Accuracy: 27.5. Result ID: 69d8412347f34ce924c04520. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Router Category":"ZS","Model":"Qwen3-30B"}. This record does not assert independent reproduction. ### Direct Accuracy: 24.9. Result ID: 69d8412347f34ce924c044fd. [Reporting paper]() Recorded evaluation configuration: {"Router Category":"Base","Model":"Qwen3-30B"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.