INDEX

Explanations

technical terms and identifiers related to software or programming contexts

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

identifiers and GUIDs

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

 enamorado

-0.53

 ansatte

-0.52

 extranjera

-0.51

 gezet

-0.51

 meisjes

-0.50

IsPostBack

-0.49

 sposa

-0.48

 boneka

-0.47

 istruzioni

-0.47

 näin

-0.46

POSITIVE LOGITS

Хьажоргаш

0.62

findpost

0.56

تقاوى

0.53

 Chwiliwch

0.52

 kasarigan

0.52

ScopeManager

0.51

RefNanny

0.49

 ComVisible

0.49

oa̍t

0.48

0.47

Activations Density 0.001%

technical terms and identifiers related to software or programming contexts

identifiers and GUIDs

No Comments

No Known Activations