INDEX

Explanations

phrases related to criticism and attacks directed towards specific individuals

oai_token-act-pair · gpt-3.5-turbo

mentions of political figures and their criticisms

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

 ILCS

-0.79

utical

-0.78

wine

-0.74

)].

-0.68

NAD

-0.67

ETF

-0.66

iton

-0.65

utic

-0.65

ospel

-0.65

<+

-0.64

POSITIVE LOGITS

 Comey

1.12

 Mexicans

1.07

 Putin

1.04

 Russia

1.01

 Pence

0.99

 Priebus

0.96

 Carrier

0.96

 Tillerson

0.95

 Corker

0.94

 Mueller

0.93

Activations Density 0.511%

phrases related to criticism and attacks directed towards specific individuals

mentions of political figures and their criticisms

No Comments

No Known Activations