INDEX

Explanations

references to man-made phenomena or items

oai_token-act-pair · gpt-3.5-turbo

phrases containing references to human-made phenomena or impacts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

byss

-0.93

alos

-0.89

osponsors

-0.85

achu

-0.82

yip

-0.82

earchers

-0.81

externalActionCode

-0.81

pload

-0.81

aucus

-0.80

asso

-0.80

POSITIVE LOGITS

 disasters

0.85

 disaster

0.79

 goodies

0.73

 horrors

0.72

 pseudo

0.71

 artificial

0.71

 medi

0.69

 goods

0.67

 interventions

0.66

 excuses

0.66

Activations Density 0.103%

references to man-made phenomena or items

phrases containing references to human-made phenomena or impacts

No Comments

No Known Activations