INDEX

Explanations

numerical information or technological terms

oai_token-act-pair · gpt-3.5-turbo

single letters or initials

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

 Rhod

-0.83

 Squirrel

-0.77

 Buffalo

-0.76

 Villa

-0.76

 Gord

-0.74

 cyan

-0.73

 Reloaded

-0.73

PB

-0.73

-0.72

rals

-0.72

POSITIVE LOGITS

is

1.63

IS

1.60

ises

1.35

ise

1.33

isin

1.27

isers

1.24

isf

1.24

isi

1.22

ISC

1.19

ISE

1.17

Activations Density 0.129%

numerical information or technological terms

single letters or initials

No Comments

No Known Activations