INDEX

Explanations

phrases related to performance metrics and statistical achievements

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

whopping, best, leading, further

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

httphttps

-0.60

 useAppContext

-0.50

 arşivlendi

-0.48

存于互联网档案馆

-0.46

 proveedores

-0.42

 समीक्षक

-0.41

 useStyles

-0.41

 procedimientos

-0.40

 veckan

-0.40

 XCTest

-0.40

POSITIVE LOGITS

 whopping

0.65

ConstraintMaker

0.63

なんと

0.56

minded

0.46

iously

0.46

addGap

0.46

高达

0.45

$\$

0.44

 staggering

0.44

わずか

0.43

Activations Density 0.018%

phrases related to performance metrics and statistical achievements

whopping, best, leading, further

No Comments

No Known Activations