INDEX

Explanations

references to physical locations or areas

oai_token-act-pair · gpt-3.5-turbo

references to areas or reasons for actions, often associated with the term "grounds."

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

 Generation

-0.70

elin

-0.65

 thru

-0.65

pocket

-0.64

 Trilogy

-0.62

icon

-0.62

 Quote

-0.62

 Reaction

-0.61

 Dynamic

-0.61

 Experiment

-0.60

POSITIVE LOGITS

 grounds

4.34

 Grounds

2.78

grounds

2.16

 ground

1.60

 premises

1.44

ground

1.36

 basis

1.34

Ground

1.31

 Ground

1.15

 reasons

1.13

Activations Density 0.016%

references to physical locations or areas

references to areas or reasons for actions, often associated with the term "grounds."

No Comments

No Known Activations