INDEX

Explanations

mentions directing the reader to additional information or resources

oai_token-act-pair · gpt-3.5-turbo

references to additional information or details

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

FontSize

-0.84

NetMessage

-0.84

pex

-0.74

soDeliveryDate

-0.74

Redditor

-0.71

hered

-0.71

avier

-0.70

aired

-0.69

xtap

-0.69

ibles

-0.68

POSITIVE LOGITS

 information

1.35

 info

1.27

 details

1.21

 insight

1.07

 INFORMATION

0.99

 detailed

0.98

 insights

0.97

 Information

0.94

 detail

0.94

 updates

0.90

Activations Density 0.039%

mentions directing the reader to additional information or resources

references to additional information or details

No Comments

No Known Activations