INDEX

Explanations

phrases related to processes and outcomes in systems

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

 loob

-0.36

 sahi

-0.35

NonNull

-0.34

 совмести

-0.33

 подо

-0.32

error

-0.32

スリ

-0.32

 Berech

-0.32

xicity

-0.31

 Erde

-0.31

POSITIVE LOGITS

 kasarigan

0.96

 suele

0.74

 suelen

0.73

 brukar

0.72

 usually

0.69

Usually

0.68

verwijspagina

0.65

 snippetHide

0.64

 Usually

0.63

 disambiguazione

0.63

Activations Density 0.186%

phrases related to processes and outcomes in systems

No Comments

No Known Activations