INDEX

Explanations

instances of abusive or inappropriate language and content moderation terms

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

offensive language and content

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

 appunto

-0.33

+:+

-0.29

すること

-0.28

 façons

-0.27

 today

-0.27

 agent

-0.27

 atas

-0.26

することで

-0.25

agent

-0.25

ComponentName

-0.25

POSITIVE LOGITS

 noDo

0.63

 الرياضيه

0.62

ApiModelProperty

0.57

fieurs

0.56

 ModelExpression

0.56

setVerticalGroup

0.56

 ujednoznacz

0.55

 FetchType

0.54

enderror

0.54

 constitutive

0.54

Activations Density 0.044%

instances of abusive or inappropriate language and content moderation terms

offensive language and content

No Comments

No Known Activations