INDEX

Explanations

expressions related to rules, guidelines, or conditions in various contexts

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

is part of

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

 poffe

-0.37

 vectorielles

-0.29

 roble

-0.29

 AssemblyProduct

-0.28

 fédé

-0.28

 LAKE

-0.27

cámara

-0.27

 parís

-0.26

 bēr

-0.26

latego

-0.26

POSITIVE LOGITS

الحياه

0.73

 &___

0.72

 CreateTagHelper

0.72

__);

0.71

ſammen

0.69

ScopeManager

0.65

хьтан

0.63

ྑ

0.63

 insuffisamment

0.61

)_/¯

0.61

Activations Density 0.104%

expressions related to rules, guidelines, or conditions in various contexts

is part of

No Comments

No Known Activations