INDEX

Explanations

content that contains offensive or inappropriate language and remarks

oai_token-act-pair · gpt-3.5-turbo

instances of derogatory or offensive language and related actions

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

anke

-0.80

 Dealer

-0.70

asio

-0.68

ellen

-0.68

elsen

-0.68

luck

-0.67

prus

-0.67

 Mechan

-0.66

CI

-0.66

 mechanism

-0.66

POSITIVE LOGITS

 slurs

1.23

 derogatory

1.18

 lewd

1.13

 blasp

1.11

 insulting

1.10

 inappropriate

1.07

 insults

1.06

 pornographic

1.05

 disrespectful

1.05

 indecent

1.03

Activations Density 0.595%

content that contains offensive or inappropriate language and remarks

instances of derogatory or offensive language and related actions

No Comments

No Known Activations