INDEX

Explanations

phrases related to fear, danger, and risks to life or safety

oai_token-act-pair · gpt-3.5-turbo

expressions related to fear for one's safety or life

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

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

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

•layer 11 - fear

No Comments

Negative Logits

âķĲ

-0.86

Ì¶

-0.86

glers

-0.72

Proxy

-0.70

oqu

-0.69

gap

-0.67

ablishment

-0.66

olon

-0.66

iterator

-0.65

Ü

-0.63

POSITIVE LOGITS

 lives

1.33

 safety

1.29

 life

1.13

 innoc

1.12

 bystanders

1.10

 wellbeing

1.09

safety

1.08

 innocent

1.04

 limb

1.04

 sanity

1.01

Activations Density 0.264%

phrases related to fear, danger, and risks to life or safety

expressions related to fear for one's safety or life

No Comments

No Known Activations