INDEX

Explanations

proper nouns related to specific entities and titles

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

names of places and things

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

featureID

-0.43

CppCodeGen

-0.35

URIComponent

-0.34

serviceWorker

-0.34

Height

-0.34

PREFERRED

-0.34

litude

-0.34

WebControls

-0.34

IntoConstraints

-0.33

ień

-0.33

POSITIVE LOGITS

 للاسماء

0.63

XmlAccessorType

0.52

 milik

0.45

 defStyleAttr

0.44

yelitis

0.44

ExtendWith

0.44

 authored

0.44

marvin

0.42

 uLocal

0.42

˾

0.42

Activations Density 0.134%

proper nouns related to specific entities and titles

names of places and things

No Comments

No Known Activations