INDEX

Explanations

phrases related to being a victim of various situations, such as crimes, disasters, or medical conditions

oai_token-act-pair · gpt-3.5-turbo

references to victims and their experiences related to various types of trauma or violence

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

heads

-0.94

head

-0.74

 volumes

-0.73

 readings

-0.73

 Levels

-0.69

prints

-0.69

liness

-0.68

houses

-0.67

ographical

-0.65

estones

-0.65

POSITIVE LOGITS

 injustice

1.09

 oppression

0.97

 tyranny

0.96

 persecution

0.93

 discrimination

0.90

 crime

0.88

 sexism

0.87

 unjust

0.85

 racism

0.85

 cruelty

0.84

Activations Density 0.149%

phrases related to being a victim of various situations, such as crimes, disasters, or medical conditions

references to victims and their experiences related to various types of trauma or violence

No Comments

No Known Activations