INDEX

Explanations

descriptors that highlight unique or distinguishing qualities

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

unique identifying feature

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

 Komunikasi

-0.49

 peines

-0.49

 Keuangan

-0.42

 enfans

-0.40

 pères

-0.40

 juges

-0.39

 hipote

-0.39

 investis

-0.39

 jueces

-0.38

 administrativo

-0.37

POSITIVE LOGITS

 distinctive

0.88

 characteristic

0.85

 Characteristic

0.83

 unmistakable

0.79

characteristic

0.77

 signature

0.75

inctive

0.75

 EconPapers

0.73

 trademark

0.73

SharedCtor

0.72

Activations Density 0.015%

descriptors that highlight unique or distinguishing qualities

unique identifying feature

No Comments

No Known Activations