INDEX

Explanations

words related to negativity or critical judgment, particularly focusing on the word "ill."

oai_token-act-pair · gpt-3.5-turbo

words associated with illness or being unwell

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

ļéĨĴ

-0.76

ICAN

-0.68

âĹ¼

-0.68

EStream

-0.68

uyomi

-0.66

£ı

-0.64

 compr

-0.64

*/(

-0.63

BOOK

-0.62

kefeller

-0.62

POSITIVE LOGITS

uminati

1.33

inois

1.17

umin

1.10

awar

1.06

nesses

1.01

ogical

1.01

iberal

1.00

ison

0.97

usive

0.95

icit

0.93

Activations Density 0.007%

words related to negativity or critical judgment, particularly focusing on the word "ill."

words associated with illness or being unwell

No Comments

No Known Activations