INDEX

Explanations

terms related to app purchases and survey statuses

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

history, completed items, or cataloged entities

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

ContentLoaded

-0.36

RegistryLite

-0.33

Alte

-0.33

Volver

-0.32

DataAnnotations

-0.32

 Weiner

-0.31

баты

-0.30

 smtplib

-0.30

 جان

-0.30

disposing

-0.30

POSITIVE LOGITS

featureID

0.71

 transfieras

0.60

 المعيارى

0.57

 nakalista

0.52

findpost

0.49

 defaultstate

0.49

oa̍t

0.48

 propOrder

0.48

 AssemblyCulture

0.48

новниш

0.48

Activations Density 0.012%

terms related to app purchases and survey statuses

history, completed items, or cataloged entities

No Comments

No Known Activations