INDEX

Explanations

locations or entities related to specific countries and states

oai_token-act-pair · gpt-3.5-turbo

mentions of specific countries and regions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

advertising

-0.78

gue

-0.76

root

-0.74

ificial

-0.73

worldly

-0.72

landish

-0.71

ylene

-0.71

gone

-0.71

icularly

-0.70

geries

-0.68

POSITIVE LOGITS

 Limited

0.96

Ltd

0.90

 Staff

0.79

 Studios

0.77

 Bureau

0.77

 Development

0.73

 Resort

0.72

 Tour

0.71

 Observatory

0.71

 Youth

0.70

Activations Density 0.157%

locations or entities related to specific countries and states

mentions of specific countries and regions

No Comments

No Known Activations