INDEX

Explanations

terms related to specific positioning or ranking

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Configuration

google/gemma-scope-2b-pt-res/layer_12/width_1m/average_l0_107

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

1,048,576

Data Type

float32

Hook Name

blocks.12.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

StoryboardSegue

-1.06

utebol

-0.69

atemala

-0.63

msgTypes

-0.60

Rhestr

-0.55

audiovisuel

-0.54

azy

-0.54

Décès

-0.54

Witness

-0.53

 propOrder

-0.52

POSITIVE LOGITS

endphp

0.58

 дописавши

0.53

 morire

0.49

 beating

0.49

HomeAsUp

0.49

 ogrodow

0.48

 peccato

0.47

avras

0.47

IntoConstraints

0.47

 spørsmål

0.46

Activations Density 0.000%

terms related to specific positioning or ranking

No Comments

No Known Activations