INDEX

Explanations

references to the ears or listening

oai_token-act-pair · gpt-3.5-turbo

terms related to audio or sound equipment

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

 Alam

-0.75

antine

-0.69

enance

-0.68

IDE

-0.65

VK

-0.64

 Gazette

-0.64

ERO

-0.63

 Vide

-0.63

 Parenthood

-0.63

 Accountability

-0.63

POSITIVE LOGITS

rings

1.50

phones

1.30

marked

1.24

ldom

1.23

wig

1.17

ring

1.12

thing

1.11

marks

1.08

phone

1.05

 lobe

1.05

Activations Density 0.039%

references to the ears or listening

terms related to audio or sound equipment

No Comments

No Known Activations