INDEX

Explanations

expressions or phrases related to technical functions or operations

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

sending response

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

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

Configuration

google/gemma-scope-9b-it-res/layer_9/width_131k/average_l0_121

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.9.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

 Scho

-0.35

 Mend

-0.33

 Fall

-0.33

ServletException

-0.33

ädchen

-0.32

 Deep

-0.32

 Read

-0.31

Gel

-0.31

もしろ

-0.31

San

-0.30

POSITIVE LOGITS

RegressionTest

0.62

 resourceCulture

0.62

enderror

0.62

 يتيمه

0.61

AndEndTag

0.59

 CreateTagHelper

0.59

 للاسماء

0.56

ScopeManager

0.55

脚注の使い方

0.55

:✨

0.54

Activations Density 0.013%

expressions or phrases related to technical functions or operations

sending response

No Comments

No Known Activations