INDEX

Explanations

the word "hot" used in various contexts

oai_token-act-pair · gpt-3.5-turbo

the term "hot" in various contexts and expressions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

ufact

-0.92

vous

-0.75

uther

-0.75

uthor

-0.74

eca

-0.71

xus

-0.70

yss

-0.69

bernatorial

-0.68

EMENT

-0.68

ĸļ

-0.67

POSITIVE LOGITS

hot

0.92

 hotter

0.91

ened

0.88

 spots

0.87

 stove

0.87

hot

0.86

 hottest

0.85

 Chili

0.85

headed

0.84

ness

0.83

Activations Density 0.013%

the word "hot" used in various contexts

the term "hot" in various contexts and expressions

No Comments

No Known Activations