INDEX

Explanations

sections that indicate legal or formal procedural elements

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

Хьажоргаш

-0.87

 незавершена

-0.83

aarrggbb

-0.82

Personendaten

-0.81

onOptions

-0.80

읗

-0.79

 lenker

-0.79

 Paglinawan

-0.78

 autorytatywna

-0.77



-0.76

POSITIVE LOGITS

0.56

 zijne

0.52

 mijne

0.48

The

0.46

 automatiques

0.44

 fırs

0.44

 myö

0.43

vė

0.42

 gouden

0.42

\{\\

0.42

Activations Density 0.013%

sections that indicate legal or formal procedural elements

No Comments

No Known Activations