INDEX

Explanations

phrases related to opening or closing actions

oai_token-act-pair · gpt-3.5-turbo

expressions related to closing and opening various concepts or entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

ĺħ

-0.65

CTV

-0.64

terday

-0.64

alter

-0.62

Zup

-0.61

verages

-0.59

ESA

-0.58

hov

-0.58

VICE

-0.58

 cele

-0.57

POSITIVE LOGITS

 doors

1.66

 door

1.45

 gates

1.33

lid

1.08

 Pandora

1.07

 mouths

1.06

 gate

1.06

 loophole

1.03

 eyes

1.02

 Door

1.02

Activations Density 0.099%

phrases related to opening or closing actions

expressions related to closing and opening various concepts or entities

No Comments

No Known Activations