INDEX

Explanations

phrases related to something being perceived as nonsensical or lacking seriousness

oai_token-act-pair · gpt-3.5-turbo

expressions of absurdity and lightheartedness

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

Reviewed

-0.84

ainer

-0.82

ioch

-0.76

yer

-0.75

affer

-0.74

rien

-0.74

amen

-0.72

âĢ¢âĢ¢âĢ¢âĢ¢

-0.71

ership

-0.69

 largeDownload

-0.68

POSITIVE LOGITS

 silly

1.18

 nonsense

0.86

 childish

0.84

 aside

0.80

pants

0.78

ishly

0.78

Pry

0.78

 crap

0.77

 stupid

0.76

 Weasley

0.74

Activations Density 0.007%

phrases related to something being perceived as nonsensical or lacking seriousness

expressions of absurdity and lightheartedness

No Comments

No Known Activations