INDEX

Explanations

references to "active" states or entities

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

active status or entities

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

Negative Logits

 harganya

-0.53

 mijne

-0.52

LookAnd

-0.48

 homenagem

-0.48

anniversaire

-0.48

urilor

-0.48

 furent

-0.47

 betweenstory

-0.47

 suaminya

-0.47

 espelho

-0.46

POSITIVE LOGITS

Active

1.14

active

1.13

 active

1.12

 Active

1.09

 ACTIVE

1.07

setActive

1.02

getActive

0.97

ACTIVE

0.92

 actives

0.90

 актив

0.89

Activations Density 0.018%

references to "active" states or entities

active status or entities

No Known Activations