INDEX

Explanations

numerical values indicating specific positions or rankings

oai_token-act-pair · gpt-3.5-turbo

numeric values related to rankings, votes, or positions

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 mosqu

-0.68

 calculator

-0.67

 legalize

-0.66

\'

-0.65

 calcul

-0.65

 sponsor

-0.65

 eleph

-0.64

 Proposition

-0.63

Muk

-0.62

 simulator

-0.62

POSITIVE LOGITS

th

2.03

rd

1.64

st

1.46

teenth

1.44

nd

1.43

ths

1.40

eenth

1.34

venth

1.20

TH

1.20

inth

1.18

Activations Density 0.044%

numerical values indicating specific positions or rankings

numeric values related to rankings, votes, or positions

No Comments

No Known Activations