INDEX

Explanations

phrases related to receiving emails with news, content updates, and promotions

oai_token-act-pair · gpt-3.5-turbo

connective words indicating relationships or additions

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

Eug

-0.59

coat

-0.57

 Toro

-0.56

ACTED

-0.54

leneck

-0.54

 Tycoon

-0.53

orah

-0.53

 Metall

-0.53

lain

-0.51

 bluff

-0.51

POSITIVE LOGITS

 promotions

0.91

 notifications

0.78

 previews

0.72

 giveaways

0.70

 invitations

0.68

Interstitial

0.68

 attachments

0.67

 warranties

0.66

 additions

0.66

 coupons

0.65

Activations Density 0.015%

phrases related to receiving emails with news, content updates, and promotions

connective words indicating relationships or additions

No Comments

No Known Activations