# Threat Keyword Inference on drug dataset DEA Intelligence Report

Source: [Threat Keyword Inference on drug dataset DEA Intelligence Report](<https://www.sota2.com/research/sota/threat-keyword-inference-on-drug-dataset-dea-intelligence-report>)

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

Data updated: 2026-02-26T06:07:21.456Z

Evaluation of model performance on Threat Keyword Inference using the DEA Intelligence Report drug dataset, measured by Top-K metrics (Top-10 to Top-50).

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 | Threat Keyword Inference |
| Dataset (including any stated split) | drug dataset DEA Intelligence Report |
| Metric | Top-10 Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 3 of 3 |
| 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

- [Top-10 Accuracy](<https://www.sota2.com/research/sota/threat-keyword-inference-on-drug-dataset-dea-intelligence-report.md?metric=Top-10%20Accuracy>): Higher is better; unit: PERCENT.

- [Top-20 Accuracy](<https://www.sota2.com/research/sota/threat-keyword-inference-on-drug-dataset-dea-intelligence-report.md?metric=Top-20%20Accuracy>): Higher is better; unit: PERCENT.

- [Top-30 Accuracy](<https://www.sota2.com/research/sota/threat-keyword-inference-on-drug-dataset-dea-intelligence-report.md?metric=Top-30%20Accuracy>): Higher is better; unit: PERCENT.

- [Top-40 Accuracy](<https://www.sota2.com/research/sota/threat-keyword-inference-on-drug-dataset-dea-intelligence-report.md?metric=Top-40%20Accuracy>): Higher is better; unit: PERCENT.

- [Top-50 Accuracy](<https://www.sota2.com/research/sota/threat-keyword-inference-on-drug-dataset-dea-intelligence-report.md?metric=Top-50%20Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Top-10 Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| DarkBERT\_CODA | 60 | Not recorded | 68fa6f741e209131cea45b44 |
| BERT\_CODA | 40 | Not recorded | 68fa6f741e209131cea45b44 |
| BERT\_Reddit | 40 | Not recorded | 68fa6f741e209131cea45b44 |

### DarkBERT\_CODA

Top-10 Accuracy: 60. Result ID: 6990c0583a6c31484902ff9a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f741e209131cea45b44.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f741e209131cea45b44.md>)

Recorded evaluation configuration: {"Base Model":"DarkBERT","Fine-tuning Corpus":"CoDA documents (drug subset)"}. This record does not assert independent reproduction.

### BERT\_CODA

Top-10 Accuracy: 40. Result ID: 6990c0583a6c31484902ff9b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f741e209131cea45b44.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa65c31e209131cea0c4ed.md>)

[Implementation](<https://github.com/google-research/bert>)

Recorded evaluation configuration: {"Base Model":"BERT","Fine-tuning Corpus":"CoDA documents (drug subset)"}. This record does not assert independent reproduction.

### BERT\_Reddit

Top-10 Accuracy: 40. Result ID: 6990c0583a6c31484902ff9c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f741e209131cea45b44.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa699b1e209131cea2308b.md>)

Recorded evaluation configuration: {"Base Model":"BERT","Fine-tuning Corpus":"Subreddit corpus (drug topic)"}. This record does not assert independent reproduction.

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