INDEX

Explanations

phrases indicating a list of items or examples

oai_token-act-pair · gpt-3.5-turbo

phrases that introduce or list examples and key ideas

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

avery

-0.89

endor

-0.88

etz

-0.87

ulhu

-0.85

iolet

-0.80

uca

-0.79

undle

-0.78

amphetamine

-0.75

byss

-0.75

culosis

-0.74

POSITIVE LOGITS

 examples

1.45

 reasons

1.30

 highlights

1.27

 excerpts

1.23

 noteworthy

1.16

 notable

1.15

 salient

1.15

 observations

1.14

 ways

1.12

 facts

1.12

Activations Density 0.102%

phrases indicating a list of items or examples

phrases that introduce or list examples and key ideas

No Comments

No Known Activations