INDEX

Explanations

references to authority and significant declarations in a religious or philosophical 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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

RTLR

-0.56

 erratic

-0.45

oluted

-0.43

rency

-0.43

الإنجليزية

-0.42

BAD

-0.42

SPATH

-0.42

duled

-0.41

 Area

-0.41

 headway

-0.41

POSITIVE LOGITS

Personendaten

0.46

ListItemIcon

0.45

+:+

0.44

脚注の使い方

0.44

 rempliss

0.42

 newItem

0.41

 تضيفلها

0.40

AISSEE

0.39

usermodel

0.38

éndolo

0.38

Activations Density 0.014%

references to authority and significant declarations in a religious or philosophical context

No Comments

No Known Activations