INDEX

Explanations

words related to names and their implications

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

 ogóle

-0.45

 imprimée

-0.40

 ejemplar

-0.40

mişti

-0.40

müş

-0.38

 encontraban

-0.36

 secar

-0.36

 antemano

-0.35

 aveug

-0.35

 exemplaires

-0.35

POSITIVE LOGITS

:✨

0.79

 незавершена

0.63

 AssemblyTitle

0.63

Tikang

0.57

RTLD

0.52

Étymologie

0.51

ByVersion

0.50

ViewImports

0.49

Inflater

0.47

XmlType

0.47

Activations Density 0.010%

words related to names and their implications

No Comments

No Known Activations