INDEX

Explanations

references to India and Indian culture

oai_token-act-pair · gpt-3.5-turbo

instances of the term "Ind" or its variations, likely referring to Indian identity or content

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Krug

-0.75

 quarry

-0.73

 quar

-0.72

 bite

-0.71

 attacker

-0.69

 semester

-0.69

 cherry

-0.66

 scrape

-0.65

Zer

-0.65

WTC

-0.65

POSITIVE LOGITS

Ind

4.00

Indian

1.87

Ind

1.84

ind

1.70

IND

1.60

Indiana

1.43

IND

1.38

Independent

1.38

 Indones

1.35

India

1.32

Activations Density 0.008%

references to India and Indian culture

instances of the term "Ind" or its variations, likely referring to Indian identity or content

No Comments

No Known Activations