INDEX

Explanations

structured data patterns or syntax

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

new client module

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

évaluateur

-0.89

 propOrder

-0.78

transQ

-0.74

 defStyleAttr

-0.68

 queſta

-0.68

$_"

-0.64

 XCTest

-0.61

RefNanny

-0.60

ChildScrollView

-0.59

ModelSerializer

-0.59

POSITIVE LOGITS

Rptr

0.41

—

0.40

langle

0.36

 their

0.36

principalColumn

0.35

the

0.34

0.32

dub

0.31

(\

0.31

 agujas

0.31

Activations Density 0.015%

structured data patterns or syntax

new client module

No Comments

No Known Activations