INDEX

Explanations

links or buttons prompting the user to click for additional or larger content

oai_token-act-pair · gpt-3.5-turbo

phrases related to image interactions and clicks

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

opoulos

-0.72

 affair

-0.68

 dispos

-0.64

-0.63

 confid

-0.63

 ducks

-0.63

 conscientious

-0.63

 tending

-0.60

 affairs

-0.59

FTWARE

-0.58

POSITIVE LOGITS

 slideshow

1.40

 enlarge

1.37

 enlarged

1.31

pdf

1.28

 enlarg

1.24

PDF

1.23

 gallery

1.20

PDF

1.14

 IMAGES

1.11

 PHOTO

1.09

Activations Density 0.138%

links or buttons prompting the user to click for additional or larger content

phrases related to image interactions and clicks

No Comments

No Known Activations