INDEX

Explanations

phrases that begin with "This" and provide information or instructions

oai_token-act-pair · gpt-3.5-turbo

demonstrative phrases indicating explanations or instructions

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

Dear

-0.78

izens

-0.73

rils

-0.72

ÃįÃį

-0.71

awed

-0.70

Ĭ±

-0.68

aths

-0.68

 Streets

-0.68

oug

-0.67

ween

-0.67

POSITIVE LOGITS

 ensures

1.49

 allows

1.42

 eliminates

1.42

 prevents

1.35

 avoids

1.35

 enables

1.33

 means

1.28

 reduces

1.27

 creates

1.26

 helps

1.25

Activations Density 0.147%

phrases that begin with "This" and provide information or instructions

demonstrative phrases indicating explanations or instructions

No Comments

No Known Activations