INDEX

Explanations

references to hypocrisy and insincerity in statements or policies

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

saying or expressing opinions

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

Mer

-0.33

 CreateTagHelper

-0.30

 sombra

-0.28

แดง

-0.27

감

-0.27

 elements

-0.27

Dav

-0.27

Sicherheit

-0.26

Đối

-0.26

toDouble

-0.26

POSITIVE LOGITS

исленность

0.68

webElementGuid

0.64

 AssemblyCompany

0.64

RefNanny

0.62

 كومونز

0.61

PreExecute

0.60

 ErrIntOverflow

0.59

OGND

0.59

RenderAtEndOf

0.58

featureID

0.57

Activations Density 0.062%

references to hypocrisy and insincerity in statements or policies

saying or expressing opinions

No Comments

No Known Activations