© Neuronpedia 2026

Privacy & Terms Blog GitHub Slack Twitter Contact

Neuronpedia

APIAssistant AxisNEW Circuit TracerNEW Steer SAE Evals Exports Community Blog Privacy & Terms Contact

GEMMA-3-4B-IT · 17-GEMMASCOPE-2-RES-16K · 1475 ｜ Neuronpedia

Home
Gemma-3-4B-IT
17-GEMMASCOPE-2-RES-16K
1475

INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Configuration

google/gemma-scope-2-4b-it/resid_post/layer_17_width_16k_l0_medium

Prompts (Dashboard)

238,145 prompts, 512 tokens each

Dataset (Dashboard)

lmsys + oasst1

No Configuration Found

Embeds

Show PlotsShow ExplanationShow ActivationsShow Test FieldShow SteerShow Link

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

𝐍

0.90

 Noise

0.88

ंटी

0.87

statuses

0.87

nit

0.87

 itulah

0.87

 Peppers

0.85

 Kabhi

0.83

 Muff

0.82

 શે

0.82

POSITIVE LOGITS

 method

1.32

 방법

1.06

方法

1.05

 manually

1.02

方式

1.01

 طريقة

0.99

 方法

0.99

 technique

0.95

 approach

0.94

方法は

0.94

Activations Density 0.569%

No Known Activations