INDEX

Explanations

phrases related to multi-use spaces or functionalities

oai_token-act-pair · gpt-3.5-turbo

references to multi-use or multi-purpose concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 Lumpur

-0.82

 Rica

-0.78

AVG

-0.75

 Cumm

-0.74

 Borders

-0.73

 Moreno

-0.73

 Akin

-0.71

 Cassidy

-0.71

 Saras

-0.70

ãĤ©

-0.70

POSITIVE LOGITS

issue

1.25

edition

1.22

dimensional

1.21

legged

1.16

dollar

1.13

exclusive

1.12

purpose

1.10

format

1.08

level

1.08

edged

1.07

Activations Density 0.048%

phrases related to multi-use spaces or functionalities

references to multi-use or multi-purpose concepts

No Comments

No Known Activations