# Long-form Factual Generation on LongFact

Source: [Long-form Factual Generation on LongFact](<https://www.sota2.com/research/sota/long-form-factual-generation-on-longfact>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-04-21T13:26:34.612Z

Evaluation of model performance on the LongFact dataset, specifically assessing Long-form Factual Generation capabilities using domain-specific Fact Recall 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 | Long-form Factual Generation |
| Dataset (including any stated split) | LongFact |
| Metric | Fact Recall (FR) - Science |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 14 of 14 |
| 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

- [Fact Recall (FR) - Science](<https://www.sota2.com/research/sota/long-form-factual-generation-on-longfact.md?metric=Fact%20Recall%20(FR)%20-%20Science>): Higher is better; unit: PERCENT.

- [Fact Recall (FR) - Technology](<https://www.sota2.com/research/sota/long-form-factual-generation-on-longfact.md?metric=Fact%20Recall%20(FR)%20-%20Technology>): Higher is better; unit: PERCENT.

- [Fact Recall (FR) - Medicine](<https://www.sota2.com/research/sota/long-form-factual-generation-on-longfact.md?metric=Fact%20Recall%20(FR)%20-%20Medicine>): Higher is better; unit: PERCENT.

- [Fact Recall (FR) - Law](<https://www.sota2.com/research/sota/long-form-factual-generation-on-longfact.md?metric=Fact%20Recall%20(FR)%20-%20Law>): Higher is better; unit: PERCENT.

- [Fact Recall (FR) - Culture](<https://www.sota2.com/research/sota/long-form-factual-generation-on-longfact.md?metric=Fact%20Recall%20(FR)%20-%20Culture>): Higher is better; unit: PERCENT.

- [Fact Recall (FR) - Event](<https://www.sota2.com/research/sota/long-form-factual-generation-on-longfact.md?metric=Fact%20Recall%20(FR)%20-%20Event>): Higher is better; unit: PERCENT.

- [Fact Recall (FR) - Communication](<https://www.sota2.com/research/sota/long-form-factual-generation-on-longfact.md?metric=Fact%20Recall%20(FR)%20-%20Communication>): Higher is better; unit: PERCENT.

- [Fact Recall (FR) - Life Sciences](<https://www.sota2.com/research/sota/long-form-factual-generation-on-longfact.md?metric=Fact%20Recall%20(FR)%20-%20Life%20Sciences>): Higher is better; unit: PERCENT.

- [Fact Recall (FR) - Average](<https://www.sota2.com/research/sota/long-form-factual-generation-on-longfact.md?metric=Fact%20Recall%20(FR)%20-%20Average>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Fact Recall (FR) - Science | Result date | Report paper |
| --- | --- | --- | --- |
| ArbGraph | 84.2 | Not recorded | 69e705247d0172140ec7f2e7 |
| MCTS-RAG | 83.8 | Not recorded | 69e705247d0172140ec7f2e7 |
| ArbGraph | 83.5 | Not recorded | 69e705247d0172140ec7f2e7 |
| MCTS-RAG | 80.4 | Not recorded | 69e705247d0172140ec7f2e7 |
| FIRE | 79.7 | Not recorded | 69e705247d0172140ec7f2e7 |
| FIRE | 79.2 | Not recorded | 69e705247d0172140ec7f2e7 |
| CRITIC | 78.4 | Not recorded | 69e705247d0172140ec7f2e7 |
| CRITIC | 78.2 | Not recorded | 69e705247d0172140ec7f2e7 |
| ReAct | 69.5 | Not recorded | 69e705247d0172140ec7f2e7 |
| CoT | 69.1 | Not recorded | 69e705247d0172140ec7f2e7 |
| CoT | 65.7 | Not recorded | 69e705247d0172140ec7f2e7 |
| ReAct | 65.6 | Not recorded | 69e705247d0172140ec7f2e7 |
| Standard RAG | 64.9 | Not recorded | 69e705247d0172140ec7f2e7 |
| Standard RAG | 59.6 | Not recorded | 69e705247d0172140ec7f2e7 |

### ArbGraph

Fact Recall (FR) - Science: 84.2. Result ID: 69e779f647f34ce924cf5dfa.

[Reporting paper](<https://www.sota2.com/research/paper/69e705247d0172140ec7f2e7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69e705247d0172140ec7f2e7.md>)

Recorded evaluation configuration: {"Backbone":"LLaMA-3.1-8B"}. This record does not assert independent reproduction.

### MCTS-RAG

Fact Recall (FR) - Science: 83.8. Result ID: 69e779f647f34ce924cf5df9.

[Reporting paper](<https://www.sota2.com/research/paper/69e705247d0172140ec7f2e7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7c001e209131cea8bcdd.md>)

Recorded evaluation configuration: {"Backbone":"LLaMA-3.1-8B"}. This record does not assert independent reproduction.

### ArbGraph

Fact Recall (FR) - Science: 83.5. Result ID: 69e779f647f34ce924cf5df3.

[Reporting paper](<https://www.sota2.com/research/paper/69e705247d0172140ec7f2e7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69e705247d0172140ec7f2e7.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-4B"}. This record does not assert independent reproduction.

### MCTS-RAG

Fact Recall (FR) - Science: 80.4. Result ID: 69e779f647f34ce924cf5df2.

[Reporting paper](<https://www.sota2.com/research/paper/69e705247d0172140ec7f2e7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7c001e209131cea8bcdd.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-4B"}. This record does not assert independent reproduction.

### FIRE

Fact Recall (FR) - Science: 79.7. Result ID: 69e779f647f34ce924cf5df1.

[Reporting paper](<https://www.sota2.com/research/paper/69e705247d0172140ec7f2e7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7c471e209131cea8d5e1.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-4B"}. This record does not assert independent reproduction.

### FIRE

Fact Recall (FR) - Science: 79.2. Result ID: 69e779f647f34ce924cf5df8.

[Reporting paper](<https://www.sota2.com/research/paper/69e705247d0172140ec7f2e7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7c471e209131cea8d5e1.md>)

Recorded evaluation configuration: {"Backbone":"LLaMA-3.1-8B"}. This record does not assert independent reproduction.

### CRITIC

Fact Recall (FR) - Science: 78.4. Result ID: 69e779f647f34ce924cf5df0.

[Reporting paper](<https://www.sota2.com/research/paper/69e705247d0172140ec7f2e7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72471e209131cea5581a.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-4B"}. This record does not assert independent reproduction.

### CRITIC

Fact Recall (FR) - Science: 78.2. Result ID: 69e779f647f34ce924cf5df7.

[Reporting paper](<https://www.sota2.com/research/paper/69e705247d0172140ec7f2e7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72471e209131cea5581a.md>)

Recorded evaluation configuration: {"Backbone":"LLaMA-3.1-8B"}. This record does not assert independent reproduction.

### ReAct

Fact Recall (FR) - Science: 69.5. Result ID: 69e779f647f34ce924cf5df6.

[Reporting paper](<https://www.sota2.com/research/paper/69e705247d0172140ec7f2e7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6eb41e209131cea41992.md>)

[Implementation](<https://react-lm.github.io/>)

Recorded evaluation configuration: {"Backbone":"LLaMA-3.1-8B"}. This record does not assert independent reproduction.

### CoT

Fact Recall (FR) - Science: 69.1. Result ID: 69e779f647f34ce924cf5df5.

[Reporting paper](<https://www.sota2.com/research/paper/69e705247d0172140ec7f2e7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e571e209131cea3f27b.md>)

Recorded evaluation configuration: {"Backbone":"LLaMA-3.1-8B"}. This record does not assert independent reproduction.

### CoT

Fact Recall (FR) - Science: 65.7. Result ID: 69e779f647f34ce924cf5def.

[Reporting paper](<https://www.sota2.com/research/paper/69e705247d0172140ec7f2e7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e571e209131cea3f27b.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-4B"}. This record does not assert independent reproduction.

### ReAct

Fact Recall (FR) - Science: 65.6. Result ID: 69e779f647f34ce924cf5dee.

[Reporting paper](<https://www.sota2.com/research/paper/69e705247d0172140ec7f2e7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6eb41e209131cea41992.md>)

[Implementation](<https://react-lm.github.io/>)

Recorded evaluation configuration: {"Backbone":"Qwen3-4B"}. This record does not assert independent reproduction.

### Standard RAG

Fact Recall (FR) - Science: 64.9. Result ID: 69e779f647f34ce924cf5ded.

[Reporting paper](<https://www.sota2.com/research/paper/69e705247d0172140ec7f2e7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69b91e209131cea23c26.md>)

[Implementation](<https://github.com/huggingface/transformers/blob/master/examples/rag/>)

Recorded evaluation configuration: {"Backbone":"Qwen3-4B"}. This record does not assert independent reproduction.

### Standard RAG

Fact Recall (FR) - Science: 59.6. Result ID: 69e779f647f34ce924cf5df4.

[Reporting paper](<https://www.sota2.com/research/paper/69e705247d0172140ec7f2e7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69b91e209131cea23c26.md>)

[Implementation](<https://github.com/huggingface/transformers/blob/master/examples/rag/>)

Recorded evaluation configuration: {"Backbone":"LLaMA-3.1-8B"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.