INDEX

Explanations

numerical comparisons and thresholds related to percentages

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

hal

-0.37

 surla

-0.37

RSpec

-0.37

mix

-0.35

msgSender

-0.34

 взять

-0.34

Cav

-0.34

ViewFeatures

-0.33

 festi

-0.32

mis

-0.31

POSITIVE LOGITS

 THAN

0.73

 than

0.71

Than

0.65

 ویکی‌پدی

0.63

 Than

0.58

than

0.58

 noDo

0.56

YOND

0.56

THAN

0.54

帖最后由

0.54

Activations Density 0.662%

numerical comparisons and thresholds related to percentages

No Comments

No Known Activations