INDEX

Explanations

mentions of individuals or entities on social media platforms

oai_token-act-pair · gpt-3.5-turbo

instances of the word "tweeted."

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

zik

-0.63

ortium

-0.63

vantage

-0.62

nea

-0.60

cised

-0.60

 immersion

-0.60

å§«

-0.60

 pedal

-0.59

por

-0.59

phal

-0.59

POSITIVE LOGITS

 tweets

0.92

"@

0.91

storms

0.91

 hasht

0.89

 hashtag

0.89

URL

0.86

Tweet

0.85

weet

0.85

 Tweet

0.83

storm

0.82

Activations Density 0.028%

mentions of individuals or entities on social media platforms

instances of the word "tweeted."

No Comments

No Known Activations