INDEX

Explanations

phrases related to needing something to be written, said, or done

oai_token-act-pair · gpt-3.5-turbo

phrases emphasizing the need for action or accountability

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Locations

-0.71

ß

-0.68

 foes

-0.68

 unden

-0.65

umerous

-0.65

VERTISEMENT

-0.63

 teasp

-0.62

ischer

-0.62

IVERS

-0.62

 readiness

-0.61

POSITIVE LOGITS

who

0.75

 whom

0.73

else

0.73

 knows

0.72

 bothered

0.72

lihood

0.71

kn

0.68

 remotely

0.67

 complaining

0.63

 legitimately

0.63

Activations Density 0.209%

phrases related to needing something to be written, said, or done

phrases emphasizing the need for action or accountability

No Comments

No Known Activations