INDEX

Explanations

phrases related to espionage and spying

oai_token-act-pair · gpt-3.5-turbo

references to espionage or spy-related activities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

esville

-0.79

à¼

-0.70

xual

-0.70

 Kurd

-0.68

 Explain

-0.66

Interstitial

-0.66

 Centauri

-0.65

clus

-0.65

urther

-0.64

imeo

-0.61

POSITIVE LOGITS

glass

0.96

ware

0.94

ionage

0.93

oleon

0.90

 satellites

0.87

 spying

0.85

CHQ

0.84

sonian

0.83

dropping

0.82

doms

0.80

Activations Density 0.051%

phrases related to espionage and spying

references to espionage or spy-related activities

No Comments

No Known Activations