INDEX

Explanations

proper nouns of different nationalities and ethnicities

oai_token-act-pair · gpt-3.5-turbo

references to specific nationalities or ethnicities

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

oral

-0.71

olate

-0.71

 headquarters

-0.65

appa

-0.64

 Kraft

-0.63

ms

-0.63

ED

-0.63

 Downtown

-0.62

UM

-0.62

ija

-0.62

POSITIVE LOGITS

 Spani

3.53

 Frenchman

2.40

 Swed

1.83

 Scots

1.66

 Californ

1.45

 Britons

1.45

 Dane

1.43

 Italians

1.43

 Scot

1.37

 Pole

1.36

Activations Density 0.047%

proper nouns of different nationalities and ethnicities

references to specific nationalities or ethnicities

No Comments

No Known Activations