INDEX

Explanations

references to programming elements and data structures

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

Swatches, Dense, App_Functions, pitch, Ex

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

@@

-0.30

 skär

-0.29

mbg

-0.29

{{--

-0.28

 brick

-0.27

▲

-0.27

«

-0.27

『

-0.27

<h3>

-0.27

“

-0.27

POSITIVE LOGITS

 للمعارف

1.00

 utafitiHapana

0.79

 مرئيه

0.75

BeginContext

0.72

awtextra

0.71

 otomatig

0.71

 مشين

0.69

exitRule

0.67

SBATCH

0.66

+#+#

0.64

Activations Density 0.039%

references to programming elements and data structures

Swatches, Dense, App_Functions, pitch, Ex

No Comments

No Known Activations