INDEX

Explanations

proper nouns

oai_token-act-pair · gpt-3.5-turbo

references to individuals and their titles or roles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 multiplying

-0.71

 Weird

-0.71

 clicking

-0.69

Translation

-0.67

 scrolling

-0.67

Researchers

-0.67

shoot

-0.66

----------------------------------------------------------------

-0.66

--------------------------------------------------------

-0.64

 Rapt

-0.64

POSITIVE LOGITS

 campaigned

1.41

 chairs

1.39

 chaired

1.36

 oversaw

1.35

 presided

1.33

 resigned

1.30

 opposes

1.25

 clashed

1.20

 ascended

1.17

 served

1.17

Activations Density 0.258%

proper nouns

references to individuals and their titles or roles

No Comments

No Known Activations