# Textual Concept Interpretability on LLaVA-NeXT textual concepts

Source: [Textual Concept Interpretability on LLaVA-NeXT textual concepts](<https://www.sota2.com/research/sota/textual-concept-interpretability-on-llava-next-textual-concepts>)

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

Data updated: 2026-06-23T08:16:42.512Z

Evaluation of model interpretability focusing on textual concepts derived from the LLaVA-NeXT dataset, measured using Detection (Det.) and Fuzziness (Fuzz.) scores.

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 | Textual Concept Interpretability |
| Dataset (including any stated split) | LLaVA-NeXT textual concepts |
| Metric | Detection Score |
| 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

- [Detection Score](<https://www.sota2.com/research/sota/textual-concept-interpretability-on-llava-next-textual-concepts.md?metric=Detection%20Score>): Higher is better; unit: PERCENT.

- [Fuzziness Score](<https://www.sota2.com/research/sota/textual-concept-interpretability-on-llava-next-textual-concepts.md?metric=Fuzziness%20Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Detection Score | Result date | Report paper |
| --- | --- | --- | --- |
| Llama 3.1 | 81 | Not recorded | 6a3a12e4206542251d7c6df1 |
| LLaVA | 78 | Not recorded | 6a3a12e4206542251d7c6df1 |
| Gemma 2 | 76 | Not recorded | 6a3a12e4206542251d7c6df1 |

### Llama 3.1

Detection Score: 81. Result ID: 6a3a40dffbc373592c34a1f1.

[Reporting paper](<https://www.sota2.com/research/paper/6a3a12e4206542251d7c6df1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75d11e209131cea6a028.md>)

[Implementation](<https://github.com/meta-llama/llama-agentic-system>)

Recorded evaluation configuration: {"Model Size":"8b"}. This record does not assert independent reproduction.

### LLaVA

Detection Score: 78. Result ID: 6a3a40dffbc373592c34a1f2.

[Reporting paper](<https://www.sota2.com/research/paper/6a3a12e4206542251d7c6df1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a3a12e4206542251d7c6df1.md>)

Recorded evaluation configuration: {"Framework":"Multidata\_SAE"}. This record does not assert independent reproduction.

### Gemma 2

Detection Score: 76. Result ID: 6a3a40dffbc373592c34a1f0.

[Reporting paper](<https://www.sota2.com/research/paper/6a3a12e4206542251d7c6df1.md>)

Recorded evaluation configuration: {"Model Size":"9b"}. This record does not assert independent reproduction.

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