INDEX

Explanations

phrases related to appearances and how things look

oai_token-act-pair · gpt-3.5-turbo

phrases indicating visual perception or appearance

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

âĢ¢âĢ¢âĢ¢âĢ¢

-0.70

RANT

-0.67

verning

-0.65

ASE

-0.63

apo

-0.62

 Demand

-0.61

hement

-0.60

 thereto

-0.59

Major

-0.59

iddler

-0.59

POSITIVE LOGITS

 looks

3.34

 Looks

2.17

 look

2.03

look

1.97

 looked

1.91

 glances

1.90

Looks

1.88

 stares

1.85

 LOOK

1.78

 smells

1.56

Activations Density 0.018%

phrases related to appearances and how things look

phrases indicating visual perception or appearance

No Comments

No Known Activations