INDEX

Explanations

derogatory terms and expressions, including personal attacks and insults

oai_token-act-pair · gpt-3.5-turbo

derogatory terms and discussions around defamation and negative speech related to individuals or groups

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 transitions

-0.76

 milestone

-0.72

 longitudinal

-0.68

 transition

-0.65

ELE

-0.65

 wearable

-0.65

 outdoor

-0.65

 infrared

-0.64

 Transition

-0.63

 outgoing

-0.61

POSITIVE LOGITS

ueless

1.06

 hypocrisy

1.04

 slander

1.03

 hypocr

1.02

 hypocritical

1.02

udicrous

1.01

entious

1.00

 bigotry

0.98

 contempt

0.96

 insulting

0.96

Activations Density 0.318%

derogatory terms and expressions, including personal attacks and insults

derogatory terms and discussions around defamation and negative speech related to individuals or groups

No Comments

No Known Activations