INDEX

Explanations

code-related 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

Affirmed

-0.44

CodedInputStream

-0.43

 medži

-0.40

 bershka

-0.36

Попис

-0.35

 durability

-0.35

 [''],

-0.35

 Durability

-0.35

 contextLoads

-0.35

 startIndex

-0.34

POSITIVE LOGITS

 autorytatywna

0.65

 Италијани

0.59

 CanadaChoose

0.53

DockStyle

0.51

 للاسماء

0.50

 defaultstate

0.50

 nahilalakip

0.49

 EconPapers

0.49

Tikang

0.48

 newBuilder

0.47

Activations Density 0.179%

code-related elements

No Comments

No Known Activations