INDEX

Explanations

elements related to specific numeric or categorical data in discussions

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

behaviors, prepared, starch, labor, drag, user

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

XmlAccessorType

-0.39

 Annahme

-0.28

EncBinary

-0.27

シル

-0.27

middels

-0.26

 Brust

-0.26

AutoScale

-0.25

 Schwer

-0.25

进去

-0.24

ById

-0.24

POSITIVE LOGITS

 Administrativna

0.79

#+#

0.77

 kasarigan

0.74

########.

0.66

Tembelea

0.64

OGND

0.62

:+:

0.60

хьтан

0.58

 Мексичка

0.58

 NSCoder

0.57

Activations Density 0.063%

elements related to specific numeric or categorical data in discussions

behaviors, prepared, starch, labor, drag, user

No Comments

No Known Activations