INDEX

Explanations

mentions of the word "American"

oai_token-act-pair · gpt-3.5-turbo

references to the term "American."

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

NB

-0.85

cffffcc

-0.80

_>

-0.76

Prev

-0.74

ÃŁ

-0.73

heed

-0.71

theless

-0.69

RH

-0.68

Redditor

-0.67

Spoiler

-0.66

POSITIVE LOGITS

 Indian

0.95

 Samoa

0.94

 American

0.91

 Americans

0.91

 Dream

0.86

 Federation

0.85

ICAN

0.84

 Heritage

0.84

 Indians

0.83

 Airlines

0.83

Activations Density 0.039%

mentions of the word "American"

references to the term "American."

No Comments

No Known Activations