INDEX

Explanations

words related to a physical structure or position

oai_token-act-pair · gpt-3.5-turbo

variations of the word "stand."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 routed

-0.67

 seeds

-0.63

 Royal

-0.61

 involved

-0.57

 colon

-0.57

 forbidden

-0.56

 redistributed

-0.56

 Hastings

-0.56

fat

-0.55

 composition

-0.55

POSITIVE LOGITS

stand

4.80

stood

2.30

Stand

2.03

standing

1.94

 stand

1.53

 Stand

1.51

 STAND

1.40

 stands

1.36

grounds

1.14

held

1.10

Activations Density 0.007%

words related to a physical structure or position

variations of the word "stand."

No Comments

No Known Activations