INDEX

Explanations

phrases that inquire about types or categories

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

 betweenstory

-0.56

 <<<<<<<<<<<<<<

-0.45

UIControlState

-0.43

mybatisplus

-0.41

 conmigo

-0.40

PyTuple

-0.40

 Ainsi

-0.39

})->

-0.39

)++;

-0.38

 prêtres

-0.37

POSITIVE LOGITS

どんな

0.65

どういう

0.59

 kinds

0.56

Serializable

0.54

mendous

0.54

什么样的

0.53

どのような

0.52

propTypes

0.51

 milyen

0.51

TypeOf

0.50

Activations Density 0.005%

phrases that inquire about types or categories

No Comments

No Known Activations