INDEX

Explanations

concepts related to feelings of shame and embarrassment

oai_token-act-pair · gpt-3.5-turbo

expressions of shame and embarrassment

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

agnetic

-0.80

ondo

-0.78

yip

-0.76

irements

-0.75

ello

-0.74

enegger

-0.72

ingham

-0.72

foreseen

-0.71

arms

-0.69

skill

-0.69

POSITIVE LOGITS

 ashamed

1.03

faced

0.83

 Shame

0.78

 Zucker

0.78

 embarrassed

0.72

 embarrassment

0.71

 shame

0.70

ness

0.70

rets

0.69

NESS

0.67

Activations Density 0.010%

concepts related to feelings of shame and embarrassment

expressions of shame and embarrassment

No Comments

No Known Activations