INDEX

Explanations

varying descriptions of different items or actions in a list

oai_token-act-pair · gpt-3.5-turbo

phrases related to unique or standout characteristics or comparisons

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

 Accountability

-0.54

 withhold

-0.50

 disclosures

-0.49

 withholding

-0.49

nesty

-0.48

 prosecutions

-0.47

]."

-0.47

 meanwhile

-0.47

).[

-0.46

 bargaining

-0.46

POSITIVE LOGITS

uin

0.54

oji

0.53

 resemblance

0.52

itamin

0.50

 resemb

0.49

 orig

0.49

ogie

0.48

 ancestor

0.48

ãĤ©

0.47

pron

0.47

Activations Density 4.045%

varying descriptions of different items or actions in a list

phrases related to unique or standout characteristics or comparisons

No Comments

No Known Activations