INDEX

Explanations

references to power laws and scaling relationships in data

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

power law and linear relations

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

 Numerade

-0.40

!*\

-0.38

 Locator

-0.37

VIAF

-0.37

횟

-0.36

cases

-0.35

 robo

-0.35

 Origin

-0.34

TagMode

-0.34

 festivals

-0.34

POSITIVE LOGITS

'{@

0.57

 vuestros

0.52

 slope

0.51

iastes

0.50

THISDAY

0.49

akterysty

0.48

 steeper

0.48

 desmotivaciones

0.47

rungsseite

0.47

 slopes

0.46

Activations Density 0.270%

references to power laws and scaling relationships in data

power law and linear relations

No Comments

No Known Activations