INDEX

Explanations

specific brand names or product components

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

places and capitalized proper names

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

transQ

-0.45

httphttps

-0.44

Cyfarwyddwr

-0.44

WithIOException

-0.42

ValueStyle

-0.39

 ostavi

-0.36

copied

-0.36

 cedo

-0.35

onomy

-0.35

lccn

-0.35

POSITIVE LOGITS

minecraftforge

0.54

AnimationsModule

0.48

WebVitals

0.47

IntoConstraints

0.46

IsPostBack

0.45

CppCodeGen

0.45

batim

0.42

URIComponent

0.41

AccessorTable

0.41

writeFieldEnd

0.41

Activations Density 0.038%

specific brand names or product components

places and capitalized proper names

No Comments

No Known Activations