INDEX

Explanations

questions and phrases indicating understanding or clarification

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

новниш

-0.64

 Italijanski

-0.63

сылкі

-0.60

ftagPool

-0.57

 للمعارف

-0.57

Билгалдахарш

-0.56

ۜ

-0.56

Handlung

-0.55

 հղումներ

-0.54

SequentialGroup

-0.54

POSITIVE LOGITS

 Abreu

0.31

 sacré

0.30

 abogados

0.29

Obfuscated

0.28

 sorella

0.28

sstil

0.28

ğunu

0.28

lewati

0.28

 lágrimas

0.28

middels

0.27

Activations Density 0.051%

questions and phrases indicating understanding or clarification

No Comments

No Known Activations