INDEX

Explanations

proper nouns

oai_token-act-pair · gpt-3.5-turbo

references to "eigenfaces" used in face recognition techniques

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

umbnail

-0.68

 Hour

-0.64

 thru

-0.61

 convincing

-0.60

 hours

-0.60

 Rockies

-0.59

 gulf

-0.59

akening

-0.59

 earnest

-0.59

 Archangel

-0.59

POSITIVE LOGITS

igen

1.49

esis

1.20

baum

0.93

stadt

0.89

omics

0.88

etics

0.88

ocide

0.85

ously

0.84

ãĥ³ãĤ¸

0.81

wald

0.81

Activations Density 0.004%

proper nouns

references to "eigenfaces" used in face recognition techniques

No Comments

No Known Activations