INDEX

Explanations

phrases indicating results, conclusions, or enumerations in a structured format

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ValueStyle

-0.67

 tartalomajánló

-0.65

hyrchwyd

-0.64

:✨

-0.62

OGND

-0.61

ंदीखरीदारी

-0.57

🇶

-0.55

 noDo

-0.54

tagHelperRunner

-0.54

ViewImports

-0.51

POSITIVE LOGITS

 following

0.59

 folgende

0.47

如下

0.47

 suivante

0.47

 folgender

0.46

 suivants

0.46

 nasled

0.44

 seguintes

0.43

 follows

0.42

 следующее

0.42

Activations Density 0.283%

phrases indicating results, conclusions, or enumerations in a structured format

No Comments

No Known Activations