INDEX

Explanations

words related to technological utility and importance

oai_token-act-pair · gpt-3.5-turbo

terms related to usefulness and quality

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

gemony

-0.62

arlane

-0.62

livion

-0.62

uthor

-0.61

everal

-0.58

roy

-0.57

Gene

-0.56

 Whitman

-0.55

Roy

-0.55

unia

-0.55

POSITIVE LOGITS

if

1.26

 because

1.12

 when

1.12

 unless

1.01

 since

1.01

 considering

0.98

 depending

0.97

for

0.96

 whenever

0.91

when

0.88

Activations Density 0.219%

words related to technological utility and importance

terms related to usefulness and quality

No Comments

No Known Activations