INDEX

Explanations

the word "OK" with a high level of activation

oai_token-act-pair · gpt-3.5-turbo

the phrase "OK" indicating acceptance or approval

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

dimension

-0.76

riber

-0.72

perial

-0.72

å¥³

-0.72

raz

-0.71

Topics

-0.69

universal

-0.68

ãĤ¼ãĤ¦ãĤ¹

-0.68

hig

-0.68

requ

-0.67

POSITIVE LOGITS

OK

1.23

lahoma

1.22

OK

1.00

AY

0.95

ok

0.87

 okay

0.78

 Okay

0.78

 Alright

0.74

 Tradable

0.73

Ok

0.72

Activations Density 0.004%

the word "OK" with a high level of activation

the phrase "OK" indicating acceptance or approval

No Comments

No Known Activations