INDEX

Explanations

phrases directing users to leave comments or interact below a given content

oai_token-act-pair · gpt-3.5-turbo

references to comments or interactions in the comment section of a document

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

Freedom

-0.71

uay

-0.70

Sense

-0.70

olly

-0.69

eg

-0.67

ggle

-0.65

CTR

-0.64

Choice

-0.64

Ring

-0.64

ãĤ¸

-0.63

POSITIVE LOGITS

 below

0.97

 veter

0.93

below

0.83

 plateau

0.81

 eleph

0.76

 tradem

0.75

estyles

0.73

hirt

0.72

 carbohyd

0.72

 crest

0.72

Activations Density 0.017%

phrases directing users to leave comments or interact below a given content

references to comments or interactions in the comment section of a document

No Comments

No Known Activations