INDEX

Explanations

pejorative and offensive language

oai_token-act-pair · gpt-3.5-turbo

derogatory insults and offensive language directed at individuals

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

ItemImage

-0.85

 departures

-0.78

Newsletter

-0.75

isSpecialOrderable

-0.74

Enlarge

-0.74

Interstitial

-0.73

 reperto

-0.70

displayText

-0.69

è£ħ

-0.69

 Installation

-0.68

POSITIVE LOGITS

 bastard

1.60

 bitch

1.60

 fuck

1.53

 shit

1.46

 dude

1.44

 asshole

1.43

 cunt

1.43

 idiot

1.40

 crap

1.38

 whore

1.37

Activations Density 0.393%

pejorative and offensive language

derogatory insults and offensive language directed at individuals

No Comments

No Known Activations