INDEX

Explanations

proper nouns in the form of names or locations

oai_token-act-pair · gpt-3.5-turbo

mentions of the name "Ke" in various contexts

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

orp

-0.77

DNA

-0.77

zbollah

-0.76

ment

-0.68

TEXT

-0.65

rant

-0.65

stration

-0.63

gypt

-0.62

MENT

-0.61

 scrimmage

-0.60

POSITIVE LOGITS

Ke

3.70

Ke

2.96

KE

1.83

Kee

1.51

ke

1.46

Ki

1.29

Ko

1.29

 Keane

1.24

Kes

1.23

ke

1.19

Activations Density 0.028%

proper nouns in the form of names or locations

mentions of the name "Ke" in various contexts

No Comments

No Known Activations