# Knowledge Conflict Resolution on KID-Bench v2 Source: [Knowledge Conflict Resolution on KID-Bench v2]() Collection index: [llms.txt]() Data updated: 2026-04-28T09:10:57.936Z Evaluation of model performance on resolving knowledge conflicts using the KID-Bench v2 dataset, measured across various difficulty levels (A, B, C-Light, C-Medium, C-Deep). 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 | Knowledge Conflict Resolution | | Dataset (including any stated split) | KID-Bench v2 | | Metric | Performance (Difficulty A) | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 4 of 4 | | 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 - [Performance (Difficulty A)](): Higher is better; unit: PERCENT. - [Performance (Difficulty B)](): Higher is better; unit: PERCENT. - [Performance (Difficulty C-Light)](): Higher is better; unit: PERCENT. - [Performance (Difficulty C-Medium)](): Higher is better; unit: PERCENT. - [Performance (Difficulty C-Deep)](): Higher is better; unit: PERCENT. ## Recorded results | Method | Performance (Difficulty A) | Result date | Report paper | | --- | --- | --- | --- | | RAG | 97.6 | Not recorded | 69f03f5255a3c32af89fa35f | | Conflict-Aware | 97.1 | Not recorded | 69f03f5255a3c32af89fa35f | | D2L Baseline | 96.7 | Not recorded | 69f03f5255a3c32af89fa35f | | SLB | 95.5 | Not recorded | 69f03f5255a3c32af89fa35f | ### RAG Performance (Difficulty A): 97.6. Result ID: 69f0795847f34ce924d69728. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Knowledge internalization strategy":"doc in context"}. This record does not assert independent reproduction. ### Conflict-Aware Performance (Difficulty A): 97.1. Result ID: 69f0795847f34ce924d6972b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Knowledge internalization strategy":"internalized into weights"}. This record does not assert independent reproduction. ### D2L Baseline Performance (Difficulty A): 96.7. Result ID: 69f0795847f34ce924d69729. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Knowledge internalization strategy":"internalized into weights"}. This record does not assert independent reproduction. ### SLB Performance (Difficulty A): 95.5. Result ID: 69f0795847f34ce924d6972a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Knowledge internalization strategy":"internalized into weights","Adaptation parameters":"25%/1.75×"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.