INDEX

Explanations

locations or directions related to the southern direction

oai_token-act-pair · gpt-3.5-turbo

references to the direction "south."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

•space_direction

south

No Comments

Negative Logits

elsen

-0.79

66666666

-0.77

ãĤ¤ãĥĪ

-0.72

ICLE

-0.72

thood

-0.70

 "$:/

-0.67

roma

-0.67

anamo

-0.67

Books

-0.65

ensity

-0.65

POSITIVE LOGITS

wark

1.06

ward

1.01

west

0.83

iors

0.81

ampton

0.78

wind

0.78

fing

0.78

bound

0.77

wards

0.75

overs

0.75

Activations Density 0.029%

locations or directions related to the southern direction

references to the direction "south."

No Comments

No Known Activations