INDEX

Explanations

phrases related to perceived dangers or threats

oai_token-act-pair · gpt-3.5-turbo

references to danger or threats to life and safety

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

Ì¶

-0.71

 Edited

-0.71

Proxy

-0.71

 Debt

-0.71

sort

-0.68

gap

-0.67

ãĥ¼ãĤ¯

-0.66

venants

-0.65

 lump

-0.64

otten

-0.63

POSITIVE LOGITS

 safety

1.24

 lives

1.21

 innocent

1.18

 bystanders

1.18

 hostages

1.14

 innoc

1.12

 civilians

1.10

safety

1.10

 pedestrians

1.07

angering

1.05

Activations Density 0.327%

phrases related to perceived dangers or threats

references to danger or threats to life and safety

No Comments

No Known Activations