INDEX

Explanations

references to the location "Quebec"

oai_token-act-pair · gpt-3.5-turbo

references to Quebec or related geographic designations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 GOODMAN

-0.78

ãĥ¼ãĥĨãĤ£

-0.77

 eleph

-0.76

ãĥŁ

-0.74

 batting

-0.72

schild

-0.72

 cuff

-0.68

ERSON

-0.67

é¾įå

-0.67

é¾įå¥ĳå£«

-0.66

POSITIVE LOGITS

zon

1.14

Que

1.02

Que

1.00

ues

0.91

bec

0.89

que

0.88

agan

0.86

erness

0.83

eg

0.83

er

0.83

Activations Density 0.005%

references to the location "Quebec"

references to Quebec or related geographic designations

No Comments

No Known Activations