INDEX

Explanations

information related to law enforcement, counter-terrorism measures, and legal penalties

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

phrases related to terrorism and legal consequences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_post

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

 bible

-0.78

rye

-0.77

vom

-0.74

 ejac

-0.73

 plent

-0.72

 paddle

-0.70

 Hatt

-0.69

 coffin

-0.69

agra

-0.68

 gorge

-0.68

POSITIVE LOGITS

 terrorist

1.14

 counterterrorism

1.12

foreign

1.11

 terrorism

1.09

intelligence

1.08

 espionage

1.08

terrorist

1.08

 nationals

1.06

Foreign

1.03

pecially

1.03

Activations Density 0.487%

information related to law enforcement, counter-terrorism measures, and legal penalties

phrases related to terrorism and legal consequences

No Comments

No Known Activations