INDEX

Explanations

words related to website menus and settings

oai_token-act-pair · gpt-3.5-turbo

references to various topics, forums, and options within a structured setting

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

lethal

-0.71

onte

-0.70

inos

-0.62

wit

-0.60

ereo

-0.60

 rigged

-0.59

weed

-0.59

ilaterally

-0.58

 strained

-0.58

authent

-0.58

POSITIVE LOGITS

 section

1.89

 page

1.89

 pages

1.63

tab

1.57

 webpage

1.52

page

1.49

 sections

1.47

 sidebar

1.42

 pane

1.41

 Page

1.41

Activations Density 0.435%

words related to website menus and settings

references to various topics, forums, and options within a structured setting

No Comments

No Known Activations