# Cyber-Threat Intelligence Attribution on LLM-CTI Attribution

Source: [Cyber-Threat Intelligence Attribution on LLM-CTI Attribution](<https://www.sota2.com/research/sota/cyber-threat-intelligence-attribution-on-llm-cti-attribution>)

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

Data updated: 2026-05-26T07:09:35.550Z

Evaluation of model performance on the LLM-CTI Attribution dataset across multiple Cyber-Threat Intelligence tasks, including Threat Actor Linking, TTP Extraction, Campaign Attribution, and Evidence Weighting.

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 | Cyber-Threat Intelligence Attribution |
| Dataset (including any stated split) | LLM-CTI Attribution |
| Metric | Threat Actor Linking Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 15 of 15 |
| 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

- [Threat Actor Linking Accuracy](<https://www.sota2.com/research/sota/cyber-threat-intelligence-attribution-on-llm-cti-attribution.md?metric=Threat%20Actor%20Linking%20Accuracy>): Higher is better; unit: PERCENT.

- [TTP Extraction F1 Score](<https://www.sota2.com/research/sota/cyber-threat-intelligence-attribution-on-llm-cti-attribution.md?metric=TTP%20Extraction%20F1%20Score>): Higher is better; unit: PERCENT.

- [Campaign Attribution Accuracy](<https://www.sota2.com/research/sota/cyber-threat-intelligence-attribution-on-llm-cti-attribution.md?metric=Campaign%20Attribution%20Accuracy>): Higher is better; unit: PERCENT.

- [Infrastructure Reuse F1 Score](<https://www.sota2.com/research/sota/cyber-threat-intelligence-attribution-on-llm-cti-attribution.md?metric=Infrastructure%20Reuse%20F1%20Score>): Higher is better; unit: PERCENT.

- [Language/Style Profiling Accuracy](<https://www.sota2.com/research/sota/cyber-threat-intelligence-attribution-on-llm-cti-attribution.md?metric=Language%2FStyle%20Profiling%20Accuracy>): Higher is better; unit: PERCENT.

- [False Flag Detection F1 Score](<https://www.sota2.com/research/sota/cyber-threat-intelligence-attribution-on-llm-cti-attribution.md?metric=False%20Flag%20Detection%20F1%20Score>): Higher is better; unit: PERCENT.

- [Evidence Weighting BLEU](<https://www.sota2.com/research/sota/cyber-threat-intelligence-attribution-on-llm-cti-attribution.md?metric=Evidence%20Weighting%20BLEU>): Higher is better; unit: PERCENT.

- [Relation Graph Building F1 Score](<https://www.sota2.com/research/sota/cyber-threat-intelligence-attribution-on-llm-cti-attribution.md?metric=Relation%20Graph%20Building%20F1%20Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Threat Actor Linking Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| G5 | 89.2 | Not recorded | 68fa7bc11e209131cea8a354 |
| Go4 | 87.1 | Not recorded | 68fa7bc11e209131cea8a354 |
| GEM | 82.2 | Not recorded | 68fa7bc11e209131cea8a354 |
| MIX | 77.3 | Not recorded | 68fa7bc11e209131cea8a354 |
| SPT | 77.1 | Not recorded | 68fa7bc11e209131cea8a354 |
| QWN | 75.3 | Not recorded | 68fa7bc11e209131cea8a354 |
| CBS | 70.7 | Not recorded | 68fa7bc11e209131cea8a354 |
| ZYS | 70.4 | Not recorded | 68fa7bc11e209131cea8a354 |
| CLD | 65.2 | Not recorded | 68fa7bc11e209131cea8a354 |
| FSC | 64.3 | Not recorded | 68fa7bc11e209131cea8a354 |
| DHT | 60.8 | Not recorded | 68fa7bc11e209131cea8a354 |
| LL70 | 59.8 | Not recorded | 68fa7bc11e209131cea8a354 |
| GRK | 52.8 | Not recorded | 68fa7bc11e209131cea8a354 |
| CB0 | 52.6 | Not recorded | 68fa7bc11e209131cea8a354 |
| LLY | 41.3 | Not recorded | 68fa7bc11e209131cea8a354 |

### G5

Threat Actor Linking Accuracy: 89.2. Result ID: 6a154714fbc373592c0c270a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc11e209131cea8a354.md>)

Recorded evaluation configuration: {"Model Category":"Industry-leading, general-purpose"}. This record does not assert independent reproduction.

### Go4

Threat Actor Linking Accuracy: 87.1. Result ID: 6a154714fbc373592c0c270b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc11e209131cea8a354.md>)

Recorded evaluation configuration: {"Model Category":"Industry-leading, general-purpose"}. This record does not assert independent reproduction.

### GEM

Threat Actor Linking Accuracy: 82.2. Result ID: 6a154714fbc373592c0c270d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc11e209131cea8a354.md>)

Recorded evaluation configuration: {"Model Category":"Industry-leading, general-purpose"}. This record does not assert independent reproduction.

### MIX

Threat Actor Linking Accuracy: 77.3. Result ID: 6a154714fbc373592c0c270f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc11e209131cea8a354.md>)

Recorded evaluation configuration: {"Model Category":"Industry-leading, general-purpose"}. This record does not assert independent reproduction.

### SPT

Threat Actor Linking Accuracy: 77.1. Result ID: 6a154714fbc373592c0c2717.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc11e209131cea8a354.md>)

Recorded evaluation configuration: {"Model Category":"Cybersecurity-specialized"}. This record does not assert independent reproduction.

### QWN

Threat Actor Linking Accuracy: 75.3. Result ID: 6a154714fbc373592c0c2710.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc11e209131cea8a354.md>)

Recorded evaluation configuration: {"Model Category":"Industry-leading, general-purpose"}. This record does not assert independent reproduction.

### CBS

Threat Actor Linking Accuracy: 70.7. Result ID: 6a154714fbc373592c0c2716.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc11e209131cea8a354.md>)

Recorded evaluation configuration: {"Model Category":"Cybersecurity-specialized"}. This record does not assert independent reproduction.

### ZYS

Threat Actor Linking Accuracy: 70.4. Result ID: 6a154714fbc373592c0c2714.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc11e209131cea8a354.md>)

Recorded evaluation configuration: {"Model Category":"Cybersecurity-specialized"}. This record does not assert independent reproduction.

### CLD

Threat Actor Linking Accuracy: 65.2. Result ID: 6a154714fbc373592c0c270c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc11e209131cea8a354.md>)

Recorded evaluation configuration: {"Model Category":"Industry-leading, general-purpose"}. This record does not assert independent reproduction.

### FSC

Threat Actor Linking Accuracy: 64.3. Result ID: 6a154714fbc373592c0c2712.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc11e209131cea8a354.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a521e209131cea82724.md>)

Recorded evaluation configuration: {"Model Category":"Cybersecurity-specialized"}. This record does not assert independent reproduction.

### DHT

Threat Actor Linking Accuracy: 60.8. Result ID: 6a154714fbc373592c0c2718.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc11e209131cea8a354.md>)

Recorded evaluation configuration: {"Model Category":"Cybersecurity-specialized"}. This record does not assert independent reproduction.

### LL70

Threat Actor Linking Accuracy: 59.8. Result ID: 6a154714fbc373592c0c270e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc11e209131cea8a354.md>)

Recorded evaluation configuration: {"Model Category":"Industry-leading, general-purpose"}. This record does not assert independent reproduction.

### GRK

Threat Actor Linking Accuracy: 52.8. Result ID: 6a154714fbc373592c0c2711.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc11e209131cea8a354.md>)

Recorded evaluation configuration: {"Model Category":"Industry-leading, general-purpose"}. This record does not assert independent reproduction.

### CB0

Threat Actor Linking Accuracy: 52.6. Result ID: 6a154714fbc373592c0c2713.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc11e209131cea8a354.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7ae51e209131cea85c05.md>)

Recorded evaluation configuration: {"Model Category":"Cybersecurity-specialized"}. This record does not assert independent reproduction.

### LLY

Threat Actor Linking Accuracy: 41.3. Result ID: 6a154714fbc373592c0c2715.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc11e209131cea8a354.md>)

Recorded evaluation configuration: {"Model Category":"Cybersecurity-specialized"}. This record does not assert independent reproduction.

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