INDEX

Explanations

mentions of geographical locations, specifically Afghanistan

oai_token-act-pair · gpt-3.5-turbo

mentions of Afghanistan and related terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

creen

-0.81

yss

-0.79

vous

-0.74

Hunt

-0.72

 Collider

-0.72

 constitu

-0.70

*/(

-0.67

ometimes

-0.66

ynt

-0.66

 philos

-0.65

POSITIVE LOGITS

istan

1.19

ghan

1.06

 Afghanistan

1.04

 Afghan

1.03

 Kabul

1.02

 Taliban

1.00

 Albania

0.96

 Afgh

0.96

 Afghans

0.95

istani

0.93

Activations Density 0.041%

mentions of geographical locations, specifically Afghanistan

mentions of Afghanistan and related terms

No Comments

No Known Activations