INDEX

Explanations

references to technical specifications and document structures, particularly in a programming context

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

sections 2-000 or IV-

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

cop

-0.35

évaluateur

-0.34

hyrchwyd

-0.34

 <>",

-0.33

 utafitiHapana

-0.32

 steaks

-0.31

 посту

-0.31

FormBuilder

-0.31

<bos>

-0.31

AppComponent

-0.31

POSITIVE LOGITS

Morfologia

0.52

 ModelExpression

0.52

CppMethod

0.52

IVEREF

0.52

værende

0.49

таратура

0.48

StoreMessageInfo

0.48

 Kebijakan

0.48

źródło

0.47

InitVars

0.46

Activations Density 0.095%

references to technical specifications and document structures, particularly in a programming context

sections 2-000 or IV-

No Comments

No Known Activations