INDEX

Explanations

keywords related to email subscriptions or news roundups

oai_token-act-pair · gpt-3.5-turbo

mentions of "roundup"

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

cession

-0.69

ippi

-0.68

oji

-0.66

bling

-0.64

Kay

-0.64

wig

-0.63

Ã±a

-0.62

 constructs

-0.62

 sunk

-0.61

essee

-0.61

POSITIVE LOGITS

 roundup

1.33

 Roundup

0.81

weekly

0.78

ãĥ¼ãĥ³

0.72

agonists

0.71

 GOODMAN

0.70

endium

0.70

 Manson

0.68

icles

0.67

ÃįÃį

0.67

Activations Density 0.007%

keywords related to email subscriptions or news roundups

mentions of "roundup"

No Comments

No Known Activations