INDEX

Explanations

references to negative traits or behaviors

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

variables and their relationships

np_acts-logits-general · gemini-2.0-flash

trait matching or variables

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 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

 duquel

-0.71

 contiennent

-0.56

 complètes

-0.56

 Normdatei

-0.55

críbete

-0.54

PYX

-0.53

 completas

-0.50

 équip

-0.49

xiety

-0.48

Strap

-0.47

POSITIVE LOGITS

 kasarigan

0.67

 مُعرِّف

0.64

AutoresizingMask

0.63

 list

0.62

 HttpNotFound

0.61

 utafitiHapana

0.60

esModule

0.58

CppCodeGen

0.58

 known

0.57

SequentialGroup

0.57

Activations Density 0.988%

references to negative traits or behaviors

variables and their relationships

trait matching or variables

No Comments

No Known Activations