INDEX

Explanations

punctuation marks that indicate the structure of sentences

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

reportWebVitals

-0.56

tagext

-0.54

 CreateTagHelper

-0.49

 protoimpl

-0.49

 noDo

-0.48

 utafitiHapana

-0.45

Jîn

-0.43

GEBURTSDATUM

-0.43

 betweenstory

-0.42

 pter

-0.41

POSITIVE LOGITS

UnusedPrivate

0.55

+#+#

0.54

 INTERESAR

0.50

\{\\

0.49

ucap

0.45

 становника

0.44

 recalling

0.43

 scattata

0.42

instancetype

0.42

 يتيمه

0.42

Activations Density 0.002%

punctuation marks that indicate the structure of sentences

No Comments

No Known Activations