INDEX

Explanations

phrases related to negative emotions, particularly shame

oai_token-act-pair · gpt-3.5-turbo

references to the concept of shame and associated emotional expressions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ersive

-0.86

enhagen

-0.84

ancial

-0.78

ichick

-0.73

IFE

-0.73

aeda

-0.73

agnetic

-0.72

agall

-0.67

ullivan

-0.67

ITNESS

-0.66

POSITIVE LOGITS

faced

1.34

fully

1.30

face

0.97

ously

0.91

ful

0.91

fulness

0.91

 shame

0.89

imaru

0.86

 judgement

0.85

 Shame

0.84

Activations Density 0.029%

phrases related to negative emotions, particularly shame

references to the concept of shame and associated emotional expressions

No Comments

No Known Activations