INDEX

Explanations

adjectives indicating excessive or exaggerated actions or qualities

oai_token-act-pair · gpt-3.5-turbo

phrases and contexts related to the concept of being "over" something or excessive behavior

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

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

 Isle

-0.71

 Ranger

-0.69

Doe

-0.68

 Heist

-0.67

 Gemini

-0.62

Eps

-0.61

FORE

-0.61

ORY

-0.59

 Dynamo

-0.59

 Independent

-0.58

POSITIVE LOGITS

priced

1.36

bearing

1.33

blown

1.33

ste

1.31

confidence

1.28

reaching

1.27

worked

1.27

reach

1.25

stated

1.25

valued

1.25

Activations Density 0.065%

adjectives indicating excessive or exaggerated actions or qualities

phrases and contexts related to the concept of being "over" something or excessive behavior

No Comments

No Known Activations