INDEX

Explanations

references to objects or activities related to bikes

oai_token-act-pair · gpt-3.5-turbo

references to bicycles

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

 Solomon

-0.72

 largeDownload

-0.70

Az

-0.69

alties

-0.69

Ter

-0.65

 Crimson

-0.64

Sev

-0.64

Pers

-0.64

ulative

-0.63

agents

-0.63

POSITIVE LOGITS

 bike

3.59

 bikes

2.94

 Bike

2.94

bike

2.83

 bicycle

2.70

 biking

2.56

 bicycles

2.22

 motorcycle

2.18

 Bicycle

2.17

 cyclist

2.11

Activations Density 0.026%

references to objects or activities related to bikes

references to bicycles

No Comments

No Known Activations