INDEX

Explanations

specific references to numerical or quantifiable data

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

turn off, disappear, become activated

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

forRoot

-0.41

 Casanova

-0.38

JvmStatic

-0.37

Dichloropropane

-0.36

épendance

-0.36

 Militär

-0.36

 dalamnya

-0.35

BindingResult

-0.35

 שוליים

-0.35

 OnInit

-0.34

POSITIVE LOGITS

DebuggerNonUser

0.44

CppMethod

0.41

crollView

0.40

وصلات

0.39

 expressly

0.38

extAlignment

0.38

IRQn

0.38

dinga

0.38

免

0.37

ffine

0.37

Activations Density 0.147%

specific references to numerical or quantifiable data

turn off, disappear, become activated

No Comments

No Known Activations