INDEX

Explanations

dates written in different languages

oai_token-act-pair · gpt-3.5-turbo

occurrences of the character sequence "ör"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 Bengal

-0.71

 Crusader

-0.68

 Virgin

-0.67

Brave

-0.65

birth

-0.65

corn

-0.65

 Cavaliers

-0.64

 Blacks

-0.63

 Mattis

-0.63

 Crus

-0.62

POSITIVE LOGITS

Ã¶r

1.22

Ã¶

1.21

andom

1.13

Ã°

1.05

Ã¼

1.03

Ã¶n

0.99

Ã¤

0.98

sson

0.96

Ã¡n

0.90

Ã©n

0.88

Activations Density 0.005%

dates written in different languages

occurrences of the character sequence "ör"

No Comments

No Known Activations