INDEX

Explanations

mathematical symbols and expressions

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

code elements and structure

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

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

AddTagHelper

-0.73

 queſta

-0.73

SuspendLayout

-0.59

expandindo

-0.58

<unused43>

-0.57

<unused16>

-0.57

<unused52>

-0.57

<unused79>

-0.57

<unused74>

-0.57

<unused3>

-0.57

POSITIVE LOGITS

the

0.63

his

0.47

its

0.46

 they

0.46

 every

0.42

帖最后由

0.41

our

0.38

who

0.37

 their

0.37

ditto

0.36

Activations Density 0.063%

mathematical symbols and expressions

code elements and structure

No Comments

No Known Activations