INDEX

Explanations

actions or events described with specific numerical values

oai_token-act-pair · gpt-3.5-turbo

actions involving folding, stacking, or organizing items

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

grounds

-0.70

AppData

-0.69

limits

-0.68

 textbooks

-0.67

 Econom

-0.66

 prisons

-0.64

 moderators

-0.62

Econom

-0.62

econom

-0.61

 warehouses

-0.61

POSITIVE LOGITS

 announcement

1.32

 arrival

1.31

 insertion

1.19

 departure

1.18

 sighting

1.18

 occurrence

1.15

 foray

1.14

 takeover

1.12

 declaration

1.11

 visit

1.10

Activations Density 0.938%

actions or events described with specific numerical values

actions involving folding, stacking, or organizing items

No Comments

No Known Activations