INDEX

Explanations

words related to hierarchy or organization, especially when structured in a sequential or ranking order

oai_token-act-pair · gpt-3.5-turbo

phrases that include the word "order" followed by numerical values or related descriptors

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

 stump

-0.68

iov

-0.67

></

-0.65

iland

-0.65

>[

-0.64

iday

-0.64

gow

-0.62

>]

-0.62

zinski

-0.61

)].

-0.60

POSITIVE LOGITS

 magnitude

1.22

 precedence

1.00

rontal

0.86

 thumb

0.75

lege

0.72

ety

0.72

 nuns

0.70

 particulars

0.70

 affairs

0.70

 Templ

0.69

Activations Density 0.050%

words related to hierarchy or organization, especially when structured in a sequential or ranking order

phrases that include the word "order" followed by numerical values or related descriptors

No Comments

No Known Activations