INDEX

Explanations

profane and offensive language

oai_token-act-pair · gpt-3.5-turbo

explicit and vulgar language

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

Completed

-0.75

INAL

-0.74

utterstock

-0.72

 guiActiveUn

-0.71

arcity

-0.70

istar

-0.69

 reluct

-0.69

PsyNetMessage

-0.68

CONCLUS

-0.66

OSP

-0.66

POSITIVE LOGITS

hole

1.03

holes

0.90

shit

0.86

face

0.83

xual

0.82

fuck

0.81

mma

0.81

manship

0.79

pants

0.77

nuts

0.76

Activations Density 0.051%

profane and offensive language

explicit and vulgar language

No Comments

No Known Activations