INDEX

Explanations

acknowledgments of contributions and support in collaborative work

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

making things possible

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.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

 HasFactory

-0.63

 MatDialog

-0.57

ectady

-0.56

 ujednoznacz

-0.56

ulemon

-0.55

 leaſt

-0.53

 kaynağından

-0.53

DPI

-0.51

 disambiguazione

-0.50

 propOrder

-0.50

POSITIVE LOGITS

 насељу

0.40

‚¬

0.37

Ty

0.32

ifer

0.31

 reaching

0.30

alsa

0.30

さまで

0.29

 reached

0.29

ditor

0.29

ậy

0.29

Activations Density 0.011%

acknowledgments of contributions and support in collaborative work

making things possible

No Comments

No Known Activations