INDEX

Explanations

the word "Kyle" at various activations, with a particularly high activation value for the stimulus "Kyle" being 10

oai_token-act-pair · gpt-3.5-turbo

the frequency of the name "Kyle" in the text

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

cens

-0.70

vernment

-0.70

fect

-0.69

 cens

-0.69

oli

-0.69

 undert

-0.68

Hung

-0.67

 instit

-0.66

 conserv

-0.65

 instrument

-0.65

POSITIVE LOGITS

 Kyle

3.46

Kyle

2.80

 Cody

1.45

 Kristen

1.30

 Brandon

1.30

 Tyler

1.28

 Grayson

1.26

 Jared

1.26

 Connor

1.26

 Kelley

1.23

Activations Density 0.014%

the word "Kyle" at various activations, with a particularly high activation value for the stimulus "Kyle" being 10

the frequency of the name "Kyle" in the text

No Comments

No Known Activations