INDEX

Explanations

elements related to programming constructs and context parameters

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

code structures

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

GEBURTSDATUM

-1.05

PreferredItem

-0.96



-0.95

 utafitiHapana

-0.94

featureID

-0.93

IUrlHelper

-0.91

 queſta

-0.91

 AssemblyTitle

-0.89

 surla

-0.88

Personendaten

-0.88

POSITIVE LOGITS

 conditions

0.30

Spec

0.28

<eos>

0.27

 koş

0.27

Nov

0.27

$\

0.25

可能

0.25

事

0.24

 condiciones

0.24

Activations Density 0.008%

elements related to programming constructs and context parameters

code structures

No Comments

No Known Activations