INDEX

Explanations

Twitter handles preceded by "@" symbol

oai_token-act-pair · gpt-3.5-turbo

mentions of social media handles or tags

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

emonium

-0.77

 pneum

-0.69

 Dise

-0.69

 circulation

-0.65

 Whitman

-0.64

quel

-0.63

 Primordial

-0.63

 defic

-0.63

 Nadu

-0.63

 ordinance

-0.62

POSITIVE LOGITS

#$

1.46

@@@@@@@@

1.11

realDonaldTrump

1.03

aic

0.93

gmail

0.89

Home

0.87

home

0.84

las

0.82

weet

0.80

GS

0.79

Activations Density 0.014%

Twitter handles preceded by "@" symbol

mentions of social media handles or tags

No Comments

No Known Activations