INDEX

Explanations

the word "native" followed by various contexts such as nationalities, languages, and peoples

oai_token-act-pair · gpt-3.5-turbo

references to native speakers or native populations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

bard

-0.88

udence

-0.80

abeth

-0.72

aday

-0.69

=-=-=-=-

-0.68

ÃĥÃĤ

-0.68

urat

-0.65

issors

-0.65

apego

-0.65

ridor

-0.64

POSITIVE LOGITS

 americ

1.00

 Advertisement

0.90

spe

0.85

izations

0.84

 Instruments

0.83

isation

0.80

born

0.79

 speakers

0.77

ities

0.76

 Hawai

0.76

Activations Density 0.031%

the word "native" followed by various contexts such as nationalities, languages, and peoples

references to native speakers or native populations

No Comments

No Known Activations