INDEX

Explanations

phrases related to images or visual representations

oai_token-act-pair · gpt-3.5-turbo

references to the term "Picture."

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

 exped

-0.72

 zeal

-0.70

une

-0.64

 acids

-0.63

 serv

-0.62

 vend

-0.61

 malf

-0.60

 libertarian

-0.60

 immunity

-0.59

loc

-0.59

POSITIVE LOGITS

 Picture

4.11

Picture

2.42

picture

2.25

 picture

2.18

 Pictures

1.97

pict

1.75

 Photo

1.67

Pict

1.62

 pictures

1.51

 PHOTO

1.49

Activations Density 0.010%

phrases related to images or visual representations

references to the term "Picture."

No Comments

No Known Activations