INDEX

Explanations

terms related to style and sophistication

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

 Crossroads

-0.49

 Russie

-0.48

 Howie

-0.45

 Cassius

-0.44

 actualité

-0.44

 Mikey

-0.44

 raio

-0.44

 what

-0.44

DataSource

-0.43

yaç

-0.42

POSITIVE LOGITS

Elegant

1.12

elegant

1.11

 elegant

1.09

 Elegant

1.05

 elegance

0.95

 elegan

0.93

 elegantly

0.85

 elegante

0.84

 Elegance

0.83

 elegancia

0.82

Activations Density 0.005%

terms related to style and sophistication

No Comments

No Known Activations