INDEX

Explanations

phrases related to geographical directions

oai_token-act-pair · gpt-3.5-turbo

geographical directions or regions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Coffin

-0.67

 decor

-0.61

 neural

-0.60

 complete

-0.60

 induct

-0.59

 Honor

-0.59

 consumption

-0.59

 actual

-0.58

 literal

-0.58

Pun

-0.57

POSITIVE LOGITS

west

4.30

east

3.14

western

2.15

south

1.89

north

1.75

West

1.65

East

1.59

outheast

1.53

heast

1.51

central

1.38

Activations Density 0.021%

phrases related to geographical directions

geographical directions or regions

No Comments

No Known Activations