INDEX

Explanations

words related to terrorism or counter-terrorism

oai_token-act-pair · gpt-3.5-turbo

frequent mentions of the word "Terror" 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 @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

çĦ

-0.82

 Yosemite

-0.74

ãģ®éŃĶ

-0.73

éĸ

-0.71

ãģ¦

-0.70

çĭ

-0.69

éĢ

-0.68

éģ

-0.67

NRS

-0.64

 Hansen

-0.63

POSITIVE LOGITS

 Terror

1.24

abad

0.98

icides

0.90

mares

0.90

Terror

0.89

icide

0.88

ãĥ³ãĤ¸

0.88

terror

0.87

nels

0.85

gey

0.84

Activations Density 0.008%

words related to terrorism or counter-terrorism

frequent mentions of the word "Terror" and related terms

No Comments

No Known Activations