# Personalized Speech-Script Generation on Personalized Deep Research Task 4 1.0 (test) Source: [Personalized Speech-Script Generation on Personalized Deep Research Task 4 1.0 (test)]() Collection index: [llms.txt]() Data updated: 2026-05-12T15:52:59.558Z Evaluation of personalized speech-script generation quality on the test split of the Personalized Deep Research Task 4 1.0 dataset, measured using metrics related to recall (R-1, R-L), metadata adherence (Met.), completeness (Comp.), readability (Read.), and specific quality checks (C.P., P.P.). 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 | Personalized Speech-Script Generation | | Dataset (including any stated split) | Personalized Deep Research Task 4 1.0 (test) | | Metric | R-1 | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 9 of 9 | | 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 - [R-1](): Higher is better; unit: PERCENT. - [R-L](): Higher is better; unit: PERCENT. - [Met. Adherence](): Higher is better; unit: PERCENT. - [Completeness (Comp.)](): Higher is better; unit: SCALAR. - [Readability (Read.)](): Higher is better; unit: SCALAR. - [Quality Check (C.P.)](): Higher is better; unit: SCALAR. - [Quality Check (P.P.)](): Higher is better; unit: SCALAR. ## Recorded results | Method | R-1 | Result date | Report paper | | --- | --- | --- | --- | | Perplexity-DR | 41.2 | Not recorded | 6a02b47fdbc836b87d824c86 | | Gemini-DR | 40.1 | Not recorded | 6a02b47fdbc836b87d824c86 | | Grok-DR | 39.5 | Not recorded | 6a02b47fdbc836b87d824c86 | | PDR | 38.84 | Not recorded | 6a02b47fdbc836b87d824c86 | | OpenAI-DR | 38.8 | Not recorded | 6a02b47fdbc836b87d824c86 | | Iterative RAG | 38.13 | Not recorded | 6a02b47fdbc836b87d824c86 | | +Search | 38.12 | Not recorded | 6a02b47fdbc836b87d824c86 | | Zero-shot | 37.99 | Not recorded | 6a02b47fdbc836b87d824c86 | | Profile Prompting | 37.24 | Not recorded | 6a02b47fdbc836b87d824c86 | ### Perplexity-DR R-1: 41.2. Result ID: 6a034c8847f34ce924e893b6. [Reporting paper]() Recorded evaluation configuration: {"category":"Deep Research Agents"}. This record does not assert independent reproduction. ### Gemini-DR R-1: 40.1. Result ID: 6a034c8847f34ce924e893b7. [Reporting paper]() Recorded evaluation configuration: {"category":"Deep Research Agents"}. This record does not assert independent reproduction. ### Grok-DR R-1: 39.5. Result ID: 6a034c8847f34ce924e893b5. [Reporting paper]() Recorded evaluation configuration: {"category":"Deep Research Agents"}. This record does not assert independent reproduction. ### PDR R-1: 38.84. Result ID: 6a034c8847f34ce924e893bb. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"category":"Personalized","backbone":"DeepSeek R1-671B"}. This record does not assert independent reproduction. ### OpenAI-DR R-1: 38.8. Result ID: 6a034c8847f34ce924e893b8. [Reporting paper]() Recorded evaluation configuration: {"category":"Deep Research Agents"}. This record does not assert independent reproduction. ### Iterative RAG R-1: 38.13. Result ID: 6a034c8847f34ce924e893ba. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"category":"Personalized","retrieval":"multi-step retrieval","backbone":"DeepSeek R1-671B"}. This record does not assert independent reproduction. ### +Search R-1: 38.12. Result ID: 6a034c8847f34ce924e893b4. [Reporting paper]() Recorded evaluation configuration: {"category":"Non-personalized","search":"public web search","backbone":"DeepSeek R1-671B"}. This record does not assert independent reproduction. ### Zero-shot R-1: 37.99. Result ID: 6a034c8847f34ce924e893b3. [Reporting paper]() Recorded evaluation configuration: {"category":"Non-personalized","backbone":"DeepSeek R1-671B"}. This record does not assert independent reproduction. ### Profile Prompting R-1: 37.24. Result ID: 6a034c8847f34ce924e893b9. [Reporting paper]() Recorded evaluation configuration: {"category":"Personalized","context":"structured user context","backbone":"DeepSeek R1-671B"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.