INDEX

Explanations

words related to criminal activities or law enforcement

oai_token-act-pair · gpt-3.5-turbo

references to crime

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

arity

-0.71

 Centauri

-0.70

atable

-0.68

hetically

-0.65

hetics

-0.65

ovo

-0.64

por

-0.64

ersed

-0.64

estine

-0.63

)=(

-0.63

POSITIVE LOGITS

 spree

1.11

fighting

0.99

fighter

0.97

fighters

0.89

 perpetrated

0.85

ously

0.85

 ridden

0.79

 enforcement

0.78

ographs

0.76

 Victim

0.75

Activations Density 0.046%

words related to criminal activities or law enforcement

references to crime

No Comments

No Known Activations