INDEX

Explanations

function definitions and related coding constructs

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

verwijspagina

-0.69

alyptus

-0.54

 يتيمه

-0.54

Filmographie

-0.52

 ویکی‌پدیا

-0.51

 Roskov

-0.50

 виправивши

-0.49

+#+

-0.49

ViewImports

-0.48

IsContent

-0.47

POSITIVE LOGITS

########.

0.43

 remise

0.39

Kur

0.39

kur

0.38

 attend

0.38

 passion

0.36

 Absch

0.36

Ob

0.36

ay

0.36

 contractor

0.36

Activations Density 0.149%

function definitions and related coding constructs

No Comments

No Known Activations