INDEX

Explanations

words related to transformation or change

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "formed" or related variations indicating creation or development

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

lyn

-0.63

 down

-0.60

 workload

-0.60

 laundry

-0.59

 discount

-0.58

 availability

-0.58

 home

-0.58

 ease

-0.57

opp

-0.57

 lore

-0.57

POSITIVE LOGITS

formed

4.40

forming

2.59

formation

2.25

forms

2.22

form

2.12

former

1.78

FORM

1.72

formance

1.26

 formed

1.19

Form

1.11

Activations Density 0.014%

words related to transformation or change

occurrences of the word "formed" or related variations indicating creation or development

No Comments

No Known Activations