INDEX

Explanations

terms related to percentages and numbers

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

æ©

-0.83

IRT

-0.69

ciating

-0.67

 Horizons

-0.65

bley

-0.65

Ì¶

-0.64

GBT

-0.63

HCR

-0.63

gency

-0.62

engeance

-0.61

POSITIVE LOGITS

(%)

0.95

 chance

0.92

 percentage

0.88

 percentages

0.86

 rate

0.84

 variance

0.84

 thereof

0.84

 Percentage

0.83

 increase

0.79

age

0.79

Activations Density 0.055%

terms related to percentages and numbers

No Comments

No Known Activations