INDEX

Explanations

medical conditions or health-related terms

oai_token-act-pair · gpt-3.5-turbo

references to processed meats and health-related risks

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ccording

-0.85

ĸļ

-0.82

Þ

-0.76

©¶æ¥µ

-0.76

senal

-0.74

isSpecial

-0.74

è¦ļéĨĴ

-0.74

ĸļå£«

-0.74

ñ

-0.73

ĪĴ

-0.72

POSITIVE LOGITS

 properly

0.53

...

0.52

 both

0.52

 deck

0.52

li

0.50

 either

0.49

 meter

0.49

 crew

0.48

 stack

0.48

 lifes

0.47

Activations Density 0.114%

medical conditions or health-related terms

references to processed meats and health-related risks

No Comments

No Known Activations