INDEX

Explanations

places or locations

oai_token-act-pair · gpt-3.5-turbo

mentions of geographic areas or locations

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

 bows

-0.73

bow

-0.72

 whist

-0.71

 Finance

-0.68

 Netflix

-0.64

 jugg

-0.64

 Freud

-0.62

 Thanksgiving

-0.62

 oath

-0.61

 sadd

-0.60

POSITIVE LOGITS

area

4.62

Area

2.49

 area

1.50

 Area

1.49

region

1.21

zone

1.05

radius

1.04

 Areas

1.01

estate

1.00

 areas

0.97

Activations Density 0.008%

places or locations

mentions of geographic areas or locations

No Comments

No Known Activations