INDEX

Explanations

phrases containing the word "working."

oai_token-act-pair · gpt-3.5-turbo

instances of the word "Working" in various contexts

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

 face

-0.74

 mids

-0.72

 chains

-0.70

mur

-0.65

asta

-0.64

fac

-0.64

tot

-0.63

 displays

-0.62

 curses

-0.62

 obliter

-0.61

POSITIVE LOGITS

 Working

3.73

Working

2.25

working

1.43

 Work

1.35

 working

1.22

Work

1.20

 Stud

1.19

 Playing

1.16

 WORK

1.15

 Collabor

1.15

Activations Density 0.018%

phrases containing the word "working."

instances of the word "Working" in various contexts

No Comments

No Known Activations