INDEX

Explanations

spam-related terms or mentions

oai_token-act-pair · gpt-3.5-turbo

terms and phrases related to spam

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

hani

-1.10

IST

-0.78

Cel

-0.71

 Borders

-0.68

 Templ

-0.66

 Patri

-0.63

 Slave

-0.62

 Syri

-0.61

 Malt

-0.61

Enlarge

-0.60

POSITIVE LOGITS

ming

1.31

inator

0.90

 spam

0.89

icide

0.87

my

0.87

mers

0.84

ulus

0.84

trap

0.81

bags

0.81

mer

0.81

Activations Density 0.009%

spam-related terms or mentions

terms and phrases related to spam

No Comments

No Known Activations