INDEX

Explanations

interrogative and directive phrases that prompt action or inquiry

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

 ModelExpression

-0.54

 ContentValues

-0.36

executeUpdate

-0.34

FormItem

-0.34

writerow

-0.33

prepareStatement

-0.33

tgz

-0.32

 propOrder

-0.31

apimachinery

-0.31

 pros

-0.31

POSITIVE LOGITS

parency

0.58

ьажоргаш

0.57

 noDo

0.50

ToScroll

0.50

BeginInit

0.50

 Betrachtung

0.49

 hypno

0.48

RTSC

0.47

ьаж

0.46

RTDA

0.46

Activations Density 0.052%

interrogative and directive phrases that prompt action or inquiry

No Comments

No Known Activations