INDEX

Explanations

expressions of strong emotions or reactions, especially negative ones

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

imum

-0.91

 Formation

-0.86

atari

-0.79

aceutical

-0.77

itivity

-0.73

omi

-0.73

ulton

-0.72

ancial

-0.69

 Attribution

-0.68

 Virtue

-0.68

POSITIVE LOGITS

 ashamed

1.59

 humiliated

1.48

 frustrated

1.47

 disgusted

1.45

 saddened

1.44

 embarrassed

1.43

 depressed

1.42

 afraid

1.39

 powerless

1.39

 confused

1.39

Activations Density 0.214%

expressions of strong emotions or reactions, especially negative ones

No Comments

No Known Activations