INDEX

Explanations

mentions of being at the top

oai_token-act-pair · gpt-3.5-turbo

references to "top" positions or rankings

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

gm

-0.71

ellow

-0.66

ija

-0.64

fw

-0.63

ouri

-0.62

itta

-0.61

selves

-0.60

ewitness

-0.60

 Parenthood

-0.58

 fiance

-0.58

POSITIVE LOGITS

most

1.06

 level

0.86

 thereof

0.82

liest

0.76

mast

0.75

 tier

0.75

side

0.75

end

0.74

loader

0.73

of

0.72

Activations Density 0.047%

mentions of being at the top

references to "top" positions or rankings

No Comments

No Known Activations