INDEX

Explanations

phrases related to facial features and injuries

oai_token-act-pair · gpt-3.5-turbo

references to facial characteristics and features

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

liga

-0.80

rak

-0.79

rift

-0.73

wright

-0.73

ouk

-0.71

 Profit

-0.67

 Krugman

-0.67

urch

-0.66

apolis

-0.65

ok

-0.65

POSITIVE LOGITS

 facial

3.75

 beard

1.64

Fac

1.48

Face

1.38

 mustache

1.38

 Face

1.38

 genital

1.32

 cran

1.28

 hairst

1.27

acial

1.24

Activations Density 0.014%

phrases related to facial features and injuries

references to facial characteristics and features

No Comments

No Known Activations