INDEX

Explanations

tweets made by individuals

oai_token-act-pair · gpt-3.5-turbo

occurrences 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 @ 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

ajor

-0.65

shall

-0.63

cised

-0.62

zik

-0.61

cum

-0.61

å£«

-0.60

 immersion

-0.59

circ

-0.58

-|

-0.58

 Reborn

-0.58

POSITIVE LOGITS

"@

0.90

 hasht

0.89

 tweets

0.89

URL

0.89

storms

0.89

 Tweet

0.87

 hashtag

0.86

Tweet

0.85

weet

0.82

 tweet

0.79

Activations Density 0.024%

tweets made by individuals

occurrences of the word "tweeted."

No Comments

No Known Activations