INDEX

Explanations

words related to strong negative emotions like hatred

oai_token-act-pair · gpt-3.5-turbo

references to hatred and animosity

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

neau

-0.81

ert

-0.71

iti

-0.70

mint

-0.69

oper

-0.68

 Logged

-0.68

leon

-0.67

 Flex

-0.67

GMT

-0.65

usc

-0.65

POSITIVE LOGITS

 hatred

3.41

 animosity

2.18

 hostility

1.98

 hate

1.96

 resentment

1.90

 dislike

1.77

 bigotry

1.77

 hateful

1.73

 intolerance

1.73

 hating

1.72

Activations Density 0.022%

words related to strong negative emotions like hatred

references to hatred and animosity

No Comments

No Known Activations