INDEX

Explanations

informal terms of address typically used among friends or peers

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

tagHelperRunner

-0.71

 arşivlendi

-0.61

enumi

-0.58

ſcher

-0.57

 onBind

-0.57

 invokingState

-0.57

帖最后由

-0.57

 kasarigan

-0.57

 HasFactory

-0.57

 المعيارى

-0.56

POSITIVE LOGITS

 dude

0.94

Dude

0.91

 Dude

0.85

 dudes

0.78

dude

0.67

guy

0.54

 fuck

0.47

 guys

0.47

man

0.46

 freak

0.45

Activations Density 0.003%

informal terms of address typically used among friends or peers

No Comments

No Known Activations