# Multi-Hop Question Answering on HotpotQA (F1 and LasJ) Source: [Multi-Hop Question Answering on HotpotQA (F1 and LasJ)]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:43:40.558Z Evaluation of model performance on the HotpotQA dataset for multi-hop question answering, specifically measured using F1 score and LasJ metrics. 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 | Multi-Hop Question Answering | | Dataset (including any stated split) | HotpotQA | | Metric | F1 Score | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 31 of 31 | | 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 - [F1 Score](): Higher is better; unit: PERCENT. - [LasJ Score](): Higher is better; unit: PERCENT. ## Recorded results | Method | F1 Score | Result date | Report paper | | --- | --- | --- | --- | | BranPO | 64.5 | Not recorded | 6982d2f6a8b01e76cc8e99ab | | BranPO | 62.7 | Not recorded | 6982d2f6a8b01e76cc8e99ab | | CARL | 62.6 | Not recorded | 69326712d6b0dd334293430e | | GRPO | 61.4 | Not recorded | 6982d2f6a8b01e76cc8e99ab | | TreeRL | 61.1 | Not recorded | 69326712d6b0dd334293430e | | BranPO | 60.9 | Not recorded | 6982d2f6a8b01e76cc8e99ab | | ARPO | 60.9 | Not recorded | 69326712d6b0dd334293430e | | CARL | 60.8 | Not recorded | 69326712d6b0dd334293430e | | GRPO | 60.7 | Not recorded | 69326712d6b0dd334293430e | | GRPO | 60.4 | Not recorded | 6982d2f6a8b01e76cc8e99ab | | BranPO | 59.8 | Not recorded | 6982d2f6a8b01e76cc8e99ab | | TreeRPO | 59.6 | Not recorded | 69326712d6b0dd334293430e | | GRPO | 59.3 | Not recorded | 69326712d6b0dd334293430e | | CARL-Lite | 59.2 | Not recorded | 69326712d6b0dd334293430e | | GRPO | 58.6 | Not recorded | 6982d2f6a8b01e76cc8e99ab | | Tree-GRPO | 58.4 | Not recorded | 6982d2f6a8b01e76cc8e99ab | | Tree-GRPO | 58.3 | Not recorded | 6982d2f6a8b01e76cc8e99ab | | GRPO | 58.1 | Not recorded | 6982d2f6a8b01e76cc8e99ab | | GiGPO | 56.9 | Not recorded | 6982d2f6a8b01e76cc8e99ab | | Tree-GRPO | 56.8 | Not recorded | 6982d2f6a8b01e76cc8e99ab | | Tree-GRPO | 56.7 | Not recorded | 6982d2f6a8b01e76cc8e99ab | | GiGPO | 56.3 | Not recorded | 6982d2f6a8b01e76cc8e99ab | | SFT | 54.9 | Not recorded | 6982d2f6a8b01e76cc8e99ab | | SFT | 54.3 | Not recorded | 6982d2f6a8b01e76cc8e99ab | | CARL | 51.9 | Not recorded | 69326712d6b0dd334293430e | | GRPO | 47 | Not recorded | 69326712d6b0dd334293430e | | CARL | 34.9 | Not recorded | 69326712d6b0dd334293430e | | GRPO | 33.7 | Not recorded | 69326712d6b0dd334293430e | | Zero-Shot | 31.9 | Not recorded | 69326712d6b0dd334293430e | | Zero-Shot | 17.4 | Not recorded | 69326712d6b0dd334293430e | | Zero-Shot | 16.4 | Not recorded | 69326712d6b0dd334293430e | ### BranPO F1 Score: 64.5. Result ID: 69964eba5698665ec4935e99. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-4B-Instruct-2507","Training interaction limit":8,"Evaluation interaction limit":16}. This record does not assert independent reproduction. ### BranPO F1 Score: 62.7. Result ID: 69964eba5698665ec4935e96. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B-Instruct","Training interaction limit":8,"Evaluation interaction limit":16}. This record does not assert independent reproduction. ### CARL F1 Score: 62.6. Result ID: 6999886d5698665ec4f9d1fa. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"4B Reasoning","Max actions":"32"}. This record does not assert independent reproduction. ### GRPO F1 Score: 61.4. Result ID: 69964eba5698665ec4935e97. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Qwen3-4B-Instruct-2507","Training interaction limit":8,"Evaluation interaction limit":16}. This record does not assert independent reproduction. ### TreeRL F1 Score: 61.1. Result ID: 6999886d5698665ec4f9d1f5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"4B Reasoning","Max actions":"10"}. This record does not assert independent reproduction. ### BranPO F1 Score: 60.9. Result ID: 69964eba5698665ec4935eba. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B-Instruct"}. This record does not assert independent reproduction. ### ARPO F1 Score: 60.9. Result ID: 6999886d5698665ec4f9d1f6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"4B Reasoning","Max actions":"10"}. This record does not assert independent reproduction. ### CARL F1 Score: 60.8. Result ID: 6999886d5698665ec4f9d1f8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"4B Reasoning","Max actions":"10"}. This record does not assert independent reproduction. ### GRPO F1 Score: 60.7. Result ID: 6999886d5698665ec4f9d1f3. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"4B Reasoning","Max actions":"10"}. This record does not assert independent reproduction. ### GRPO F1 Score: 60.4. Result ID: 69964eba5698665ec4935e94. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B-Instruct","Training interaction limit":8,"Evaluation interaction limit":16}. This record does not assert independent reproduction. ### BranPO F1 Score: 59.8. Result ID: 69964eba5698665ec4935ebf. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-4B-Instruct-2507"}. This record does not assert independent reproduction. ### TreeRPO F1 Score: 59.6. Result ID: 6999886d5698665ec4f9d1f4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"4B Reasoning","Max actions":"10"}. This record does not assert independent reproduction. ### GRPO F1 Score: 59.3. Result ID: 6999886d5698665ec4f9d1f9. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"4B Reasoning","Max actions":"32"}. This record does not assert independent reproduction. ### CARL-Lite F1 Score: 59.2. Result ID: 6999886d5698665ec4f9d1f7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"4B Reasoning","Max actions":"10"}. This record does not assert independent reproduction. ### GRPO F1 Score: 58.6. Result ID: 69964eba5698665ec4935eb7. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B-Instruct"}. This record does not assert independent reproduction. ### Tree-GRPO F1 Score: 58.4. Result ID: 69964eba5698665ec4935e98. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-4B-Instruct-2507","Training interaction limit":8,"Evaluation interaction limit":16}. This record does not assert independent reproduction. ### Tree-GRPO F1 Score: 58.3. Result ID: 69964eba5698665ec4935e95. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B-Instruct","Training interaction limit":8,"Evaluation interaction limit":16}. This record does not assert independent reproduction. ### GRPO F1 Score: 58.1. Result ID: 69964eba5698665ec4935ebc. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Qwen3-4B-Instruct-2507"}. This record does not assert independent reproduction. ### GiGPO F1 Score: 56.9. Result ID: 69964eba5698665ec4935eb8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B-Instruct"}. This record does not assert independent reproduction. ### Tree-GRPO F1 Score: 56.8. Result ID: 69964eba5698665ec4935ebe. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-4B-Instruct-2507"}. This record does not assert independent reproduction. ### Tree-GRPO F1 Score: 56.7. Result ID: 69964eba5698665ec4935eb9. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B-Instruct"}. This record does not assert independent reproduction. ### GiGPO F1 Score: 56.3. Result ID: 69964eba5698665ec4935ebd. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-4B-Instruct-2507"}. This record does not assert independent reproduction. ### SFT F1 Score: 54.9. Result ID: 69964eba5698665ec4935ebb. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-4B-Instruct-2507"}. This record does not assert independent reproduction. ### SFT F1 Score: 54.3. Result ID: 69964eba5698665ec4935eb6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B-Instruct"}. This record does not assert independent reproduction. ### CARL F1 Score: 51.9. Result ID: 6999886d5698665ec4f9d1f1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"7B Non-Reasoning","Max actions":"32"}. This record does not assert independent reproduction. ### GRPO F1 Score: 47. Result ID: 6999886d5698665ec4f9d1f0. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"7B Non-Reasoning","Max actions":"32"}. This record does not assert independent reproduction. ### CARL F1 Score: 34.9. Result ID: 6999886d5698665ec4f9d1ee. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"3B Non-Reasoning","Max actions":"32"}. This record does not assert independent reproduction. ### GRPO F1 Score: 33.7. Result ID: 6999886d5698665ec4f9d1ed. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"3B Non-Reasoning","Max actions":"32"}. This record does not assert independent reproduction. ### Zero-Shot F1 Score: 31.9. Result ID: 6999886d5698665ec4f9d1f2. [Reporting paper]() Recorded evaluation configuration: {"Model":"4B Reasoning","Max actions":"10"}. This record does not assert independent reproduction. ### Zero-Shot F1 Score: 17.4. Result ID: 6999886d5698665ec4f9d1ef. [Reporting paper]() Recorded evaluation configuration: {"Model":"7B Non-Reasoning","Max actions":"32"}. This record does not assert independent reproduction. ### Zero-Shot F1 Score: 16.4. Result ID: 6999886d5698665ec4f9d1ec. [Reporting paper]() Recorded evaluation configuration: {"Model":"3B Non-Reasoning","Max actions":"32"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.