INDEX

Explanations

references to terrorist activities or threats

oai_token-act-pair · gpt-3.5-turbo

references to terrorism and related activities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

dit

-0.86

galitarian

-0.78

laus

-0.74

shire

-0.74

ðĿ

-0.71

 Minotaur

-0.70

erick

-0.69

ilde

-0.67

equality

-0.66

LECT

-0.66

POSITIVE LOGITS

 mastermind

1.11

 bombings

1.11

 attacks

1.08

 terrorist

1.06

 terrorists

1.05

abad

1.04

 bombing

1.04

 attack

0.97

 bombers

0.97

 threats

0.93

Activations Density 0.074%

references to terrorist activities or threats

references to terrorism and related activities

No Comments

No Known Activations