INDEX

Explanations

phrases related to giving commands or directives

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

ģĸ

-0.69

-0.67

 Combined

-0.67

ļéĨĴ

-0.64

 purported

-0.61

 ancest

-0.59

features

-0.58

 culminating

-0.58

 Vendor

-0.58

 strikingly

-0.58

POSITIVE LOGITS

 yourselves

1.45

 yourself

1.10

thy

1.00

 your

0.94

ye

0.93

ya

0.92

 Yourself

0.89

 thou

0.88

me

0.86

 fuckin

0.84

Activations Density 0.281%

phrases related to giving commands or directives

No Comments

No Known Activations