INDEX

Explanations

elements related to the publication and release history of works

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

originally released or published

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

 JpaRepository

-0.42

KURZBESCHREIBUNG

-0.42

Autoritní

-0.41

AndEndTag

-0.38

onOptions

-0.37

Backend

-0.37

 MediatR

-0.37

 precin

-0.37

correre

-0.37

 intStringLen

-0.37

POSITIVE LOGITS

 originalmente

0.47

 propOrder

0.45

 originally

0.43

proven

0.41

originally

0.40

Originally

0.40

 ursprünglich

0.39

 AssemblyVersion

0.39

 AssemblyProduct

0.39

iseite

0.38

Activations Density 0.034%

elements related to the publication and release history of works

originally released or published

No Known Activations