INDEX

Explanations

instructions or steps related to programming or technical tasks

oai_token-act-pair · gpt-3.5-turbo

questions and prompts related to instructions or actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 mund

-0.66

balls

-0.62

 jewels

-0.62

 Tasman

-0.61

 Hyder

-0.60

cedented

-0.59

Hart

-0.59

 enthusi

-0.59

cases

-0.57

 originating

-0.57

POSITIVE LOGITS

 remove

1.32

 send

1.29

 declare

1.28

 carve

1.28

 execute

1.28

 analyze

1.27

cknow

1.27

 erase

1.26

 create

1.26

 Create

1.24

Activations Density 0.452%

instructions or steps related to programming or technical tasks

questions and prompts related to instructions or actions

No Comments

No Known Activations