INDEX

Explanations

political and news-related terms and phrases

oai_token-act-pair · gpt-3.5-turbo

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

xual

-0.72

ones

-0.67

TEXTURE

-0.67

-0.65

tions

-0.64

:-

-0.64

AAF

-0.62

 Anon

-0.61

lihood

-0.60

naires

-0.59

POSITIVE LOGITS

 smartest

0.66

 celeb

0.65

anooga

0.62

Inside

0.61

 digest

0.59

inion

0.58

 Kavanaugh

0.57

 noon

0.57

 uncover

0.57

akespeare

0.57

Activations Density 0.112%

political and news-related terms and phrases

No Comments

No Known Activations