INDEX

Explanations

references to a hierarchical structure or tree-like organization

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

"root" or "Root" (case-insensitive)

np_acts-logits-general · gemini-2.0-flash

root object element node

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

PMailer

-0.66

 autorytatywna

-0.61

ValueStyle

-0.58

MIDDLEWARE

-0.56

 Ellington

-0.53

 للمعارف

-0.52

PageModule

-0.52

icolor

-0.51

Zal

-0.51

 BigNumber

-0.51

POSITIVE LOGITS

Root

1.28

root

1.23

 root

1.22

 Root

1.18

 ROOT

1.09

ROOT

1.09

getRoot

0.97

 raíz

0.88

 roots

0.87

roots

0.85

Activations Density 0.101%

references to a hierarchical structure or tree-like organization

"root" or "Root" (case-insensitive)

root object element node

No Comments

No Known Activations