INDEX

Explanations

words related to discrimination and social injustice

oai_token-act-pair · gpt-3.5-turbo

references to large entities or concepts, particularly those that are deemed negative or oppressive

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

IUM

-0.67

ability

-0.66

 Emin

-0.65

 largeDownload

-0.64

 Guest

-0.63

ICLE

-0.61

ILA

-0.61

ajor

-0.59

lihood

-0.59

 Technician

-0.59

POSITIVE LOGITS

oted

1.42

ots

1.24

wig

1.21

gest

1.18

ger

1.15

gie

1.13

amy

1.11

picture

1.07

 bucks

0.97

bang

0.96

Activations Density 0.066%

words related to discrimination and social injustice

references to large entities or concepts, particularly those that are deemed negative or oppressive

No Comments

No Known Activations