INDEX

Explanations

words related to intense criticism or anger

oai_token-act-pair · gpt-3.5-turbo

words related to betrayal or negative portrayals of individuals

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

perture

-0.74

 drill

-0.74

print

-0.69

 fries

-0.68

 Maker

-0.66

¦

-0.66

 timeout

-0.66

arations

-0.63

frog

-0.62

arent

-0.62

POSITIVE LOGITS

iven

1.01

sky

0.88

ski

0.83

ival

0.81

rov

0.81

ven

0.78

 venom

0.76

egal

0.74

gil

0.74

ulner

0.74

Activations Density 0.028%

words related to intense criticism or anger

words related to betrayal or negative portrayals of individuals

No Comments

No Known Activations