INDEX

Explanations

phrases indicating direction or purpose

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.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

 another

-0.50

 stället

-0.42

another

-0.41

 instead

-0.40

enumi

-0.40

将

-0.40

 istället

-0.39

 expenditure

-0.37

 inuti

-0.37

 Manfaat

-0.36

POSITIVE LOGITS

 CreateTagHelper

0.80

tagHelperRunner

0.71

tvguidetime

0.70

 Soluble

0.66

XmlAccessType

0.61

RenderAtEndOf

0.59

 betweenstory

0.58

ſelben

0.57

cemic

0.57

transQ

0.56

Activations Density 0.414%

phrases indicating direction or purpose

No Comments

No Known Activations