INDEX

Explanations

combinations of symbols and characters that appear in mathematical or coding contexts

oai_token-act-pair · gpt-3.5-turbo

special characters and symbols within the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

creen

-0.73

 Borders

-0.73

kens

-0.69

 Mutual

-0.66

Pie

-0.65

ktop

-0.64

 Jenner

-0.64

 externalToEVAOnly

-0.62

 Thin

-0.60

 sclerosis

-0.60

POSITIVE LOGITS

ģ

1.50

ĥ

1.44

ī

1.42

Ģ

1.41

Ĥ

1.32

ĭ

1.24

ĳ

1.23

ĩ

1.23

ı

1.20

Ĩ

1.18

Activations Density 0.004%

combinations of symbols and characters that appear in mathematical or coding contexts

special characters and symbols within the text

No Comments

No Known Activations