INDEX

Explanations

terms related to political extremism, specifically neo-Nazis

oai_token-act-pair · gpt-3.5-turbo

mentions of neo-Nazi groups and related terms

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

 hereby

-0.83

 SHARES

-0.73

 emoji

-0.72

 bets

-0.70

 affecting

-0.68

ado

-0.66

channelAvailability

-0.66

 warrants

-0.65

 overdue

-0.64

 indemn

-0.64

POSITIVE LOGITS

liberal

1.29

Nazi

1.26

Nazis

1.22

colonial

1.18

natal

1.15

fascist

1.11

traditional

1.08

social

1.05

conservative

1.04

liber

1.03

Activations Density 0.045%

terms related to political extremism, specifically neo-Nazis

mentions of neo-Nazi groups and related terms

No Comments

No Known Activations