INDEX

Explanations

questions starting with "Who"

oai_token-act-pair · gpt-3.5-turbo

questions beginning with "Who."

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

MER

-0.95

interstitial

-0.85

 Hyde

-0.71

PORT

-0.71

pit

-0.66

England

-0.64

BACK

-0.64

bal

-0.63

UTION

-0.62

mun

-0.62

POSITIVE LOGITS

soever

1.08

oping

0.91

ever

0.89

abouts

0.89

 else

0.85

 cares

0.81

ispers

0.78

tymology

0.76

ortium

0.75

 accomp

0.75

Activations Density 0.069%

questions starting with "Who"

questions beginning with "Who."

No Comments

No Known Activations