INDEX

Explanations

keywords related to instructions or processes

oai_token-act-pair · gpt-3.5-turbo

instances of the word "work" along with its various contextual applications

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ilings

-0.83

antha

-0.80

ylon

-0.79

 Flavoring

-0.74

anamo

-0.72

idium

-0.72

ModLoader

-0.70

iren

-0.66

xual

-0.64

fty

-0.64

POSITIVE LOGITS

flows

1.08

bench

1.04

station

1.03

manship

0.94

heet

0.89

hops

0.89

 ethic

0.86

river

0.83

horse

0.82

work

0.80

Activations Density 0.080%

keywords related to instructions or processes

instances of the word "work" along with its various contextual applications

No Comments

No Known Activations