INDEX

Explanations

phrases related to contradiction or contrast

oai_token-act-pair · gpt-3.5-turbo

negative phrases that indicate a lack of clarity or uncertainty

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

ä»£

-0.60

æ°

-0.57

ife

-0.55

ãĥĨ

-0.53

emale

-0.49

culosis

-0.48

obo

-0.48

å£«

-0.48

ãĤ©

-0.47

otype

-0.47

POSITIVE LOGITS

etheless

0.92

 nonetheless

0.81

 disclaim

0.65

 nevertheless

0.62

 caution

0.60

lur

0.59

 balk

0.59

 quir

0.59

 dogged

0.56

 caveats

0.54

Activations Density 2.138%

phrases related to contradiction or contrast

negative phrases that indicate a lack of clarity or uncertainty

No Comments

No Known Activations