INDEX

Explanations

instances where the word "scale" is used in a context that suggests a large or significant magnitude

oai_token-act-pair · gpt-3.5-turbo

references to magnitude or extent in various contexts

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

esson

-0.73

hiro

-0.70

uala

-0.69

jong

-0.67

unal

-0.66

pron

-0.66

nor

-0.66

ieu

-0.66

sbm

-0.66

algia

-0.66

POSITIVE LOGITS

 scale

1.11

 scales

0.95

 Scale

0.78

 replica

0.76

 invari

0.76

scale

0.76

itized

0.73

 insign

0.69

 gesture

0.69

 numer

0.66

Activations Density 0.010%

instances where the word "scale" is used in a context that suggests a large or significant magnitude

references to magnitude or extent in various contexts

No Comments

No Known Activations