INDEX

Explanations

lowercase letter "c" with high activation values

oai_token-act-pair · gpt-3.5-turbo

the character "c" 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 @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

ĪĴ

-0.70

 Yamaha

-0.69

 contagious

-0.69

 Gors

-0.65

 Maker

-0.64

 Siberia

-0.62

 CLICK

-0.62

 LAST

-0.61

 basket

-0.58

 mornings

-0.57

POSITIVE LOGITS

redits

1.32

ursor

1.26

ologne

1.25

ourses

1.24

ascade

1.22

ouple

1.21

ategories

1.21

rossover

1.20

risis

1.20

ursed

1.19

Activations Density 0.039%

lowercase letter "c" with high activation values

the character "c" in various contexts

No Comments

No Known Activations