INDEX

Explanations

words related to social media handles and usernames

oai_token-act-pair · gpt-3.5-turbo

words or abbreviations that indicate social media and communication platforms

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

SHIP

-0.75

CPC

-0.72

OPLE

-0.66

cius

-0.61

staking

-0.58

 goats

-0.57

auga

-0.55

acters

-0.54

 confines

-0.54

 agitation

-0.54

POSITIVE LOGITS

Wan

0.89

edo

0.79

acea

0.72

ratulations

0.72

leck

0.70

eda

0.69

wald

0.69

aqu

0.68

uct

0.67

Sense

0.66

Activations Density 0.136%

words related to social media handles and usernames

words or abbreviations that indicate social media and communication platforms

No Comments

No Known Activations