INDEX

Explanations

Afghan-related terms and names

oai_token-act-pair · gpt-3.5-turbo

mentions of Afghanistan and its related entities

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Livingston

-0.74

canon

-0.70

SIGN

-0.69

 refere

-0.65

 Pelicans

-0.64

clips

-0.64

Sac

-0.64

isc

-0.63

Dod

-0.62

ck

-0.61

POSITIVE LOGITS

 Afghan

3.76

 Afghans

3.23

 Kabul

2.80

 Afgh

2.68

 Afghanistan

2.61

 Taliban

2.27

 Pakistani

2.07

Af

2.03

 Yemeni

2.02

 Uzbek

1.98

Activations Density 0.033%

Afghan-related terms and names

mentions of Afghanistan and its related entities

No Comments

No Known Activations