INDEX

Explanations

phrases related to software features or functionalities that allow users to customize or manipulate aspects of a program

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate the capability or functionality of an application or feature

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

bage

-0.69

 Beasts

-0.69

 Languages

-0.66

querque

-0.65

Zen

-0.63

venant

-0.62

acci

-0.62

ItemThumbnailImage

-0.61

 Purg

-0.61

 millenn

-0.60

POSITIVE LOGITS

tered

1.14

tering

1.02

us

0.81

isters

0.80

icia

0.78

eries

0.78

ories

0.74

selves

0.73

enne

0.70

terness

0.67

Activations Density 0.023%

phrases related to software features or functionalities that allow users to customize or manipulate aspects of a program

phrases that indicate the capability or functionality of an application or feature

No Comments

No Known Activations