INDEX

Explanations

percentage values

oai_token-act-pair · gpt-3.5-turbo

numerical data related to statistics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 Kafka

-0.67

 Arabian

-0.67

 dive

-0.65

 Dickinson

-0.65

 drama

-0.63

 Zoro

-0.63

 flirt

-0.63

 deduction

-0.62

 Illum

-0.62

 theatrical

-0.62

POSITIVE LOGITS

thirds

1.47

enough

1.34

years

1.29

months

1.26

percent

1.21

million

1.18

month

1.17

times

1.14

billion

1.14

year

1.14

Activations Density 0.120%

percentage values

numerical data related to statistics

No Comments

No Known Activations