INDEX

Explanations

words related to opposition or contradiction

oai_token-act-pair · gpt-3.5-turbo

terms related to counter-arguments and counterproductive actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

©¶æ¥µ

-0.71

ogo

-0.63

é¾įå

-0.63

 Bonds

-0.62

 Vide

-0.61

 FANTASY

-0.61

weeney

-0.61

ahime

-0.60

 Likes

-0.60

 livest

-0.59

POSITIVE LOGITS

measures

0.80

dict

0.78

attack

0.76

xual

0.75

intuitive

0.75

arya

0.73

ctive

0.72

atives

0.71

argument

0.71

rad

0.71

Activations Density 0.078%

words related to opposition or contradiction

terms related to counter-arguments and counterproductive actions

No Comments

No Known Activations