INDEX

Explanations

phrases related to warning, loss, and action

oai_token-act-pair · gpt-3.5-turbo

concerns related to user experience and potential issues in communication or service

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 foremost

-0.67

ãĤ´ãĥ³

-0.66

ĸļ

-0.63

 preceded

-0.62

 conclud

-0.60

 predec

-0.59

 deem

-0.58

wise

-0.58

efully

-0.58

hler

-0.58

POSITIVE LOGITS

 cookies

0.79

 boobs

0.74

 pics

0.74

 bitcoins

0.72

 pussy

0.70

 stuff

0.69

tx

0.68

 pizza

0.68

 your

0.68

 ______

0.67

Activations Density 0.982%

phrases related to warning, loss, and action

concerns related to user experience and potential issues in communication or service

No Comments

No Known Activations