INDEX

Explanations

political language and terms

oai_token-act-pair · gpt-3.5-turbo

references to atheism and political ideologies

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

availability

-0.78

ĸļ

-0.71

Production

-0.68

 expire

-0.62

 Delivery

-0.62

 Carbuncle

-0.60

pection

-0.59

TTL

-0.57

 Cinderella

-0.56

 liner

-0.56

POSITIVE LOGITS

 extremist

1.20

 Libertarian

1.15

 libertarian

1.15

Conservative

1.14

 nationalist

1.08

 militant

1.05

odox

1.05

 denouncing

1.04

 conservatives

1.04

 evangelical

1.04

Activations Density 2.284%

political language and terms

references to atheism and political ideologies

No Comments

No Known Activations