INDEX

Explanations

words related to expressing opinions or making statements

oai_token-act-pair · gpt-3.5-turbo

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

ãĥ¯ãĥ³

-0.76

naires

-0.71

ãĤµ

-0.70

ļéĨĴ

-0.68

psc

-0.68

isode

-0.68

ITNESS

-0.63

ranged

-0.61

iatus

-0.61

ipl

-0.60

POSITIVE LOGITS

 tempting

1.27

 unclear

1.22

 undeniable

1.21

 unsur

1.20

 doubtful

1.19

 noteworthy

1.16

 abundantly

1.14

 understandable

1.14

 clear

1.10

 conceivable

1.10

Activations Density 0.128%

words related to expressing opinions or making statements

No Comments

No Known Activations