INDEX

Explanations

mentions of the location "Oregon"

oai_token-act-pair · gpt-3.5-turbo

mentions of the state of Oregon

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

 Sheikh

-0.78

 Jinn

-0.75

 Ùħ

-0.73

 lashes

-0.73

sab

-0.72

 swords

-0.72

 centrif

-0.71

fy

-0.71

sword

-0.71

 Kham

-0.70

POSITIVE LOGITS

 Oregon

3.65

Oregon

3.31

Portland

2.22

 Portland

2.15

 Idaho

2.13

 Maine

2.03

 Arizona

1.87

 Montana

1.81

 Ohio

1.81

 Utah

1.80

Activations Density 0.017%

mentions of the location "Oregon"

mentions of the state of Oregon

No Comments

No Known Activations