INDEX

Explanations

words related to scaling or proportion

oai_token-act-pair · gpt-3.5-turbo

references to scaling and the concept of scalability

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

odor

-0.80

 Lauder

-0.77

onom

-0.73

 Scotia

-0.70

was

-0.68

gamer

-0.68

present

-0.67

 Guilty

-0.67

tery

-0.67

ran

-0.67

POSITIVE LOGITS

 scaled

1.24

 scaling

0.94

 eleph

0.93

GoldMagikarp

0.88

 carbohyd

0.86

anchester

0.83

 practition

0.83

 exting

0.79

 shenan

0.78

CLASSIFIED

0.77

Activations Density 0.010%

words related to scaling or proportion

references to scaling and the concept of scalability

No Comments

No Known Activations