INDEX

Explanations

phrases related to electrical components or stores

oai_token-act-pair · gpt-3.5-turbo

references to various types of outlets or sources

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

heed

-0.76

aah

-0.75

ohan

-0.71

illusion

-0.68

 Griffith

-0.67

enforcement

-0.65

etime

-0.65

eric

-0.64

ummies

-0.63

 Feather

-0.62

POSITIVE LOGITS

 outlet

1.41

 outlets

1.12

swick

0.88

="#

0.80

icago

0.71

Accessory

0.70

ption

0.69

hub

0.69

 valve

0.69

stadt

0.69

Activations Density 0.005%

phrases related to electrical components or stores

references to various types of outlets or sources

No Comments

No Known Activations