INDEX

Explanations

numerical data formatted in a structured way

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114

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

 мәкал

-0.56

 ostavi

-0.56

 esternos

-0.55

 Paglinawan

-0.54

исленность

-0.52

 Exactos

-0.51

 propOrder

-0.51

Autoritní

-0.49

 صوتيه

-0.48

 Roskov

-0.46

POSITIVE LOGITS

 betweenstory

0.59

 liitty

0.48

 kaupung

0.48

transQ

0.47

 Khusus

0.45

Hair

0.45

Batch

0.44

Finance

0.43

Work

0.43

 Manfaat

0.43

Activations Density 0.104%

numerical data formatted in a structured way

No Comments

No Known Activations