INDEX

Explanations

dependent clauses or phrases that emphasize the importance or relevance of a previous point

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

this lest

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

 Pitcher

-0.60

 Photocase

-0.52

-0.51

Iva

-0.51

tbs

-0.50

㌘

-0.48

 Winfield

-0.48

eriks

-0.48

 Player

-0.47

 boxes

-0.47

POSITIVE LOGITS

GEBURTSDATUM

0.48

userdetails

0.41

SharedDtor

0.40

 ProtoMessage

0.40

notifyListeners

0.40

 Derfor

0.38

 Daarom

0.38

lihatan

0.37

 mattered

0.37

 ustedes

0.36

Activations Density 0.039%

dependent clauses or phrases that emphasize the importance or relevance of a previous point

this lest

No Comments

No Known Activations