INDEX

Explanations

programming constructs related to decision-making, particularly switch statements

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_33/width_131k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.33.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

ViewImports

-0.39

Pyx

-0.38

 DBNull

-0.37

toPromise

-0.36

verständlich

-0.36

 TextAppearance

-0.35

ngrx

-0.35

EndContext

-0.34

 ceiling

-0.33

LookAnd

-0.33

POSITIVE LOGITS

Diwedd

0.47

geſ

0.46

Ӕ

0.46

MigrationBuilder

0.45

 ब्रेकडाउन

0.45

 CanadaChoose

0.44

Erstellt

0.44

ukone

0.44

Kjelder

0.43

TryDecodeAsNil

0.42

Activations Density 0.008%

programming constructs related to decision-making, particularly switch statements

No Comments

No Known Activations