INDEX

Explanations

capitalized names indicating a specific individual

oai_token-act-pair · gpt-3.5-turbo

end-of-text tokens

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 unpre

-0.67

 Crimean

-0.62

cers

-0.61

 cardio

-0.61

 Glory

-0.60

 outgoing

-0.59

 minded

-0.59

 Staples

-0.57

sers

-0.57

 Kuro

-0.56

POSITIVE LOGITS

ombie

1.51

ombies

1.48

ERO

1.39

ebra

1.28

odiac

1.27

imbabwe

1.27

ooming

1.16

oom

1.14

ealous

1.11

hao

1.09

Activations Density 0.028%

capitalized names indicating a specific individual

end-of-text tokens

No Comments

No Known Activations