INDEX

Explanations

phrases or words related to various attacks or conflicts

oai_token-act-pair · gpt-3.5-turbo

references to attacks in geopolitical contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

çīĪ

-0.80

stable

-0.77

âķĲ

-0.77

nces

-0.76

igans

-0.76

ffen

-0.74

baum

-0.71

UX

-0.70

natureconservancy

-0.70

herty

-0.69

POSITIVE LOGITS

 violence

0.83

 terrorism

0.83

 extremism

0.79

 jihad

0.78

 spree

0.78

 Killing

0.77

 Brigade

0.74

 rampage

0.74

 raids

0.73

 bombing

0.72

Activations Density 0.433%

phrases or words related to various attacks or conflicts

references to attacks in geopolitical contexts

No Comments

No Known Activations