INDEX

Explanations

mentions of sweet food items

oai_token-act-pair · gpt-3.5-turbo

references to sweetness

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

Downloadha

-1.03

 Administ

-0.75

 Homeless

-0.71

Buk

-0.71

ctica

-0.71

 Occupations

-0.69

 Shutterstock

-0.68

pora

-0.68

 Statistical

-0.67

POSITIVE LOGITS

 sweet

1.36

eners

1.20

sweet

1.20

heart

1.12

ened

1.09

bread

0.97

 sweetness

0.97

itionally

0.95

ening

0.94

est

0.92

Activations Density 0.008%

mentions of sweet food items

references to sweetness

No Comments

No Known Activations