INDEX

Explanations

phrases indicating communication or action towards a specific person or organization

oai_token-act-pair · gpt-3.5-turbo

actions related to communication or inquiries made by various individuals or organizations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 Hits

-0.63

rab

-0.59

 Used

-0.59

 trailing

-0.59

 dominated

-0.57

nut

-0.55

 TOTAL

-0.55

 ranges

-0.55

 Companion

-0.55

 Dangerous

-0.55

POSITIVE LOGITS

 apologise

1.29

ask

1.29

 clarify

1.29

 discuss

1.29

 apologize

1.26

 complain

1.24

 reassure

1.20

 congratulate

1.18

 warn

1.16

 inquire

1.16

Activations Density 0.128%

phrases indicating communication or action towards a specific person or organization

actions related to communication or inquiries made by various individuals or organizations

No Comments

No Known Activations