INDEX

Explanations

phrases related to verifying human identity

oai_token-act-pair · gpt-3.5-turbo

phrases associated with user verification and consent

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 miscarriage

-0.63

ãĥ¼ãĤ¯

-0.57

 melting

-0.56

 miscar

-0.53

pport

-0.53

panic

-0.52

SPONSORED

-0.52

 Wonderful

-0.51

 gall

-0.50

 drift

-0.50

POSITIVE LOGITS

're

0.71

urrency

0.68

reth

0.65

Sign

0.65

urrencies

0.64

amera

0.63

worthiness

0.60

auga

0.60

 Sign

0.60

rame

0.59

Activations Density 0.017%

phrases related to verifying human identity

phrases associated with user verification and consent

No Comments

No Known Activations