INDEX

Explanations

reports or information related to specific individuals or organizations

oai_token-act-pair · gpt-3.5-turbo

instances of the word "reportedly."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

cest

-0.72

gers

-0.66

uit

-0.65

bern

-0.64

lap

-0.64

ger

-0.63

antics

-0.63

arium

-0.63

eer

-0.62

eem

-0.62

POSITIVE LOGITS

 located

0.77

Ĥİ

0.75

 obtained

0.73

 priced

0.71

 reportedly

0.68

 complied

0.67

 purchased

0.67

è¦ļéĨĴ

0.66

 supplied

0.66

 operated

0.65

Activations Density 0.019%

reports or information related to specific individuals or organizations

instances of the word "reportedly."

No Comments

No Known Activations