INDEX

Explanations

the word "only" with a focus on emphasizing limitations or exclusivity

oai_token-act-pair · gpt-3.5-turbo

phrases expressing the idea of something being partially complete or insufficient

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ros

-0.74

ducers

-0.66

idon

-0.63

rigan

-0.63

wealth

-0.62

rosis

-0.61

rote

-0.61

insula

-0.60

atana

-0.60

hement

-0.59

POSITIVE LOGITS

 marginally

1.08

 kidding

0.79

 scratched

0.78

 temporary

0.73

 partially

0.72

 scratching

0.68

 allowed

0.67

 accessible

0.67

 temporarily

0.66

 indirectly

0.66

Activations Density 0.064%

the word "only" with a focus on emphasizing limitations or exclusivity

phrases expressing the idea of something being partially complete or insufficient

No Comments

No Known Activations