INDEX

Explanations

mentions of reporting or contacting a specific entity, likely related to crime or incidents

oai_token-act-pair · gpt-3.5-turbo

references to crime stoppers and reporting crimes

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

 Cath

-0.73

 TRUMP

-0.72

description

-0.71

rum

-0.67

ships

-0.66

 Lazarus

-0.66

ripp

-0.63

estation

-0.62

 CONS

-0.62

scape

-0.62

POSITIVE LOGITS

oppers

1.36

ĸļ

0.96

oshenko

0.88

ogle

0.85

chedel

0.83

agascar

0.78

coni

0.78

uckle

0.78

axy

0.78

terday

0.78

Activations Density 0.016%

mentions of reporting or contacting a specific entity, likely related to crime or incidents

references to crime stoppers and reporting crimes

No Known Activations