INDEX

Explanations

references to espionage or surveillance activities

oai_token-act-pair · gpt-3.5-turbo

references to espionage and spying activities

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

xual

-0.77

ŃĶ

-0.73

opathic

-0.70

 Explain

-0.68

esville

-0.65

usable

-0.65

bal

-0.64

Course

-0.63

oses

-0.63

nick

-0.63

POSITIVE LOGITS

ionage

0.93

sonian

0.88

 spying

0.86

glass

0.81

dropping

0.80

doms

0.79

 abroad

0.78

 spoof

0.71

OSH

0.71

afia

0.71

Activations Density 0.021%

references to espionage or surveillance activities

references to espionage and spying activities

No Comments

No Known Activations