INDEX

Explanations

mentions of a specific person named "Cho."

oai_token-act-pair · gpt-3.5-turbo

references to a person named Cho

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

issance

-0.79

imov

-0.74

DragonMagazine

-0.73

 Metroid

-0.72

hyd

-0.72

aunder

-0.69

 substrate

-0.69

ior

-0.68

etime

-0.68

igue

-0.67

POSITIVE LOGITS

Cho

3.81

Cho

2.61

cho

2.17

 Choi

1.72

CHO

1.60

 Chow

1.55

CHO

1.47

Kw

1.25

 choir

1.17

cho

1.17

Activations Density 0.025%

mentions of a specific person named "Cho."

references to a person named Cho

No Comments

No Known Activations