INDEX

Explanations

phrases related to official statements or declarations being made

oai_token-act-pair · gpt-3.5-turbo

instances of direct speech or quotations

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

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

ennial

-0.78

NetMessage

-0.77

EStream

-0.75

âĦ¢:

-0.75

zon

-0.74

\.

-0.73

 disenfranch

-0.70

à¼

-0.68

Jews

-0.65

otin

-0.65

POSITIVE LOGITS

 spokesman

1.73

 spokeswoman

1.69

 spokesperson

1.68

 spokes

1.33

 officials

1.06

 director

1.05

 manager

1.05

 Superintendent

1.04

 ministry

1.03

 statement

1.03

Activations Density 0.246%

phrases related to official statements or declarations being made

instances of direct speech or quotations

No Comments

No Known Activations