INDEX

Explanations

expressions related to historical context

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

uxxxx

-0.52

sistors

-0.41

 COORDIN

-0.39

iligt

-0.38

impôt

-0.37

Vidite

-0.37

fromnode

-0.37

 Sebastián

-0.37

énom

-0.37

 Schrö

-0.37

POSITIVE LOGITS

 يتيمه

0.59

parsedMessage

0.58

Cyfeiriadau

0.56

TagMode

0.55

CloseOperation

0.55

 ویکی‌پدیا

0.52

Tembelea

0.50

SharedDtor

0.50

الحياه

0.48

MemoryWarning

0.48

Activations Density 0.045%

expressions related to historical context

No Comments

No Known Activations