INDEX

Explanations

formatting and structural elements typically used in academic or technical documents

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

section and subsection headings

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

 AssemblyCompany

-0.87

 pinulongan

-0.77

 насељу

-0.76

ImageContext

-0.75

 OFDb

-0.75

 ControllerBase

-0.74

 queſta

-0.73

TargetException

-0.73

Geplaatst

-0.71

richTextPanel

-0.71

POSITIVE LOGITS

subsection

0.31

SequentialGroup

0.26

 Thai

0.25

sprecher

0.24

 sociala

0.24

 November

0.23

 ennek

0.23

↵

0.23

 August

0.22

hdessä

0.22

Activations Density 0.001%

formatting and structural elements typically used in academic or technical documents

section and subsection headings

No Comments

No Known Activations