# Long-context conversation memory reasoning on LongMemEval-s oracle variant HuggingFace xiaowu0162 Source: [Long-context conversation memory reasoning on LongMemEval-s oracle variant HuggingFace xiaowu0162]() Collection index: [llms.txt]() Data updated: 2026-04-21T13:33:57.066Z Evaluation of model performance on long-context conversation memory reasoning using the specific LongMemEval-s oracle variant from HuggingFace (xiaowu0162), reporting detailed metrics across single-session, multi-session, temporal reasoning, and knowledge update capabilities. 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 | Long-context conversation memory reasoning | | Dataset (including any stated split) | LongMemEval-s oracle variant HuggingFace xiaowu0162 | | Metric | Single-session Accuracy (User) | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 6 of 6 | | 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 - [Single-session Accuracy (User)](): Higher is better; unit: PERCENT. - [Single-session Accuracy (Assistant)](): Higher is better; unit: PERCENT. - [Single-session Preference Score](): Higher is better; unit: PERCENT. - [Knowledge Update Accuracy](): Higher is better; unit: PERCENT. - [Temporal Reasoning Accuracy](): Higher is better; unit: PERCENT. - [Multi-session Reasoning Accuracy](): Higher is better; unit: PERCENT. - [Overall Accuracy](): Higher is better; unit: PERCENT. - [Task-Averaged Accuracy](): Higher is better; unit: PERCENT. ## Recorded results | Method | Single-session Accuracy (User) | Result date | Report paper | | --- | --- | --- | --- | | Hydra DB | 100 | Not recorded | 69e7052d7d0172140ec7f35b | | WorldDB | 98.57 | Not recorded | 69e7052d7d0172140ec7f35b | | Supermemory | 98.57 | Not recorded | 69e7052d7d0172140ec7f35b | | Zep | 92.9 | Not recorded | 69e7052d7d0172140ec7f35b | | Full-ctx | 81.4 | Not recorded | 69e7052d7d0172140ec7f35b | | Mem0 | 38.71 | Not recorded | 69e7052d7d0172140ec7f35b | ### Hydra DB Single-session Accuracy (User): 100. Result ID: 69e77cae47f34ce924cf703f. [Reporting paper]() Recorded evaluation configuration: {"LLM Backbone":"Gemini 3.0 Pro"}. This record does not assert independent reproduction. ### WorldDB Single-session Accuracy (User): 98.57. Result ID: 69e77cae47f34ce924cf703e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Answer Model":"Claude Opus 4.7","Judge Model":"Claude Sonnet 4.6","Extraction Model":"Claude Haiku 4.5","Embedding Model":"OpenAI text-embedding-3-small (1536-dim)","Retrieval K":"50","Ingestion Protocol":"session-by-session"}. This record does not assert independent reproduction. ### Supermemory Single-session Accuracy (User): 98.57. Result ID: 69e77cae47f34ce924cf7040. [Reporting paper]() Recorded evaluation configuration: {"LLM Backbone":"Gemini 3.0 Pro"}. This record does not assert independent reproduction. ### Zep Single-session Accuracy (User): 92.9. Result ID: 69e77cae47f34ce924cf7041. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LLM Backbone":"GPT-4o"}. This record does not assert independent reproduction. ### Full-ctx Single-session Accuracy (User): 81.4. Result ID: 69e77cae47f34ce924cf7042. [Reporting paper]() Recorded evaluation configuration: {"LLM Backbone":"GPT-4o"}. This record does not assert independent reproduction. ### Mem0 Single-session Accuracy (User): 38.71. Result ID: 69e77cae47f34ce924cf7043. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.