INDEX

Explanations

technical terms and phrases relevant to scientific or engineering contexts

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

conventional illustration

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

 Majefty

-0.54

AnchorStyles

-0.53

الحياه

-0.53

RTEX

-0.52

ArgumentParser

-0.52

 дописавши

-0.51

__(/*!

-0.50

jsdelivr

-0.49

astify

-0.49

ukone

-0.49

POSITIVE LOGITS

 Whereas

0.39

 wasteful

0.37

 monotonous

0.36

DeleteMapping

0.35

 NSCoder

0.35

ICIAS

0.35

 sayur

0.35

 pobreza

0.35

 pourtant

0.35

 tuttavia

0.34

Activations Density 0.109%

technical terms and phrases relevant to scientific or engineering contexts

conventional illustration

No Comments

No Known Activations