INDEX

Explanations

instructions or advice for writing content that is clear and concise

oai_token-act-pair · gpt-3.5-turbo

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

EStream

-0.76

 awoke

-0.75

 Ferdinand

-0.73

romeda

-0.72

 awoken

-0.69

ilon

-0.67

 Rutherford

-0.66

uries

-0.65

anus

-0.64

 Daly

-0.64

POSITIVE LOGITS

 descriptive

1.50

 concise

1.35

 keywords

1.26

 URLs

1.21

 formatting

1.21

 catchy

1.20

 ital

1.20

 keyword

1.17

URL

1.15

 paragraphs

1.14

Activations Density 0.505%

instructions or advice for writing content that is clear and concise

No Comments

No Known Activations