INDEX

Explanations

references to thumbs or thumb-related actions

oai_token-act-pair · gpt-3.5-turbo

references to the phrase "thumbs up" or related gestures indicating approval or disapproval

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

hammad

-0.83

Ø§ÙĦ

-0.77

RT

-0.75

 Occupations

-0.74

Downloadha

-0.72

ulkan

-0.71

mingham

-0.71

effic

-0.70

assetsadobe

-0.69

agher

-0.67

POSITIVE LOGITS

 thumb

1.21

wheel

0.86

hole

0.86

 finger

0.85

tip

0.82

eenth

0.81

pad

0.80

 fingers

0.79

 stroke

0.77

toe

0.77

Activations Density 0.006%

references to thumbs or thumb-related actions

references to the phrase "thumbs up" or related gestures indicating approval or disapproval

No Comments

No Known Activations