INDEX

Explanations

elements related to categorization and comparison

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

describing origin or type

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

withIdentifier

-0.38

 "..\..\

-0.36

ctico

-0.33

⋮

-0.31

 iconTwitter

-0.31

 resourceCulture

-0.30

題

-0.30

いざ

-0.30

WithEmail

-0.30

woff

-0.29

POSITIVE LOGITS

 AssemblyCompany

0.59

.*")]

0.58

hyrchwyd

0.53

 local

0.52

 native

0.52

 EconPapers

0.51

EndInit

0.51

地元

0.49

Rptr

0.48

pondi

0.47

Activations Density 0.078%

elements related to categorization and comparison

describing origin or type

No Comments

No Known Activations