INDEX

Explanations

terms related to criminal activities or investigations

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 @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

)=(

-0.82

ersed

-0.75

arity

-0.74

 Centauri

-0.73

por

-0.73

pread

-0.73

achev

-0.71

hetically

-0.70

galitarian

-0.69

chell

-0.68

POSITIVE LOGITS

 spree

1.04

fighting

0.86

 synd

0.85

fighter

0.83

 unfocusedRange

0.80

 perpetrated

0.79

 punishable

0.75

ously

0.74

 offenses

0.72

 crimes

0.72

Activations Density 0.036%

terms related to criminal activities or investigations

references to crime

No Comments

No Known Activations