INDEX

Explanations

mentions of high-ranking government officials and organizations related to intelligence and security

oai_token-act-pair · gpt-3.5-turbo

references to individuals in positions of authority within intelligence and military agencies

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

 Jagu

-0.86

 neighbours

-0.84

 relegation

-0.83

pel

-0.80

 terr

-0.78

 Romantic

-0.76

 mould

-0.76

 corrid

-0.75

 pubs

-0.74

cul

-0.74

POSITIVE LOGITS

 Clapper

1.35

FBI

1.27

 Pentagon

1.22

 Mattis

1.21

CIA

1.17

CIA

1.16

 Comey

1.14

 briefed

1.13

 Tillerson

1.13

 counterterrorism

1.11

Activations Density 0.907%

mentions of high-ranking government officials and organizations related to intelligence and security

references to individuals in positions of authority within intelligence and military agencies

No Comments

No Known Activations