INDEX

Explanations

specific characters or character sequences that match a particular pattern

oai_token-act-pair · gpt-3.5-turbo

special characters and symbols typically used in non-English scripts or encodings

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

 Heath

-0.72

SPONSORED

-0.67

 Starr

-0.66

 Daly

-0.66

 Hyde

-0.65

 Hastings

-0.63

ORED

-0.62

Faw

-0.62

 Barton

-0.61

 Worldwide

-0.61

POSITIVE LOGITS

Ð

1.49

Ŀ

1.31

±

1.19

ļ

1.19

Ķ

1.18

¹

1.17

ł

1.15

ĳ

1.13

Ĺ

1.13

Ł

1.12

Activations Density 0.004%

specific characters or character sequences that match a particular pattern

special characters and symbols typically used in non-English scripts or encodings

No Comments

No Known Activations