INDEX

Explanations

numerical values or count-related information

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_33/width_131k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.33.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

LabelTagHelper

-0.36

déric

-0.35

 betweenstory

-0.34

ázka

-0.29

 Lengkap

-0.29

staat

-0.27

AntiForgeryToken

-0.27

acamata

-0.26

declare

-0.26

fin

-0.26

POSITIVE LOGITS

esgue

0.68

SequentialGroup

0.64

脚注の使い方

0.63

tvguidetime

0.63

 noDo

0.60

 мәкал

0.60

WriteLiteral

0.58

tagHelperRunner

0.58

تقاوى

0.58

 Wikimedijinoj

0.55

Activations Density 0.002%

numerical values or count-related information

No Comments

No Known Activations