INDEX

Explanations

references to court cases and legal terms

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

Code, numbers, or abbreviations after periods or hyphens

np_acts-logits-general · gemini-2.0-flash

multi-language technical terms

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 22-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_22/width_16k/average_l0_72

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.22.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

 twenty

-0.99

-0.95

 Twenty

-0.92

Twenty

-0.82

twenty

-0.80

 TWENTY

-0.76

二十

-0.71

 vinte

-0.71

 nineteen

-0.65

 vingt

-0.61

POSITIVE LOGITS

Autoritní

0.74

 autorytatywna

0.63

featureID

0.58

 nahilalakip

0.57

 <<<<<<<<<<<<<<

0.53

 trois

0.53

DMETHOD

0.51

Composable

0.51

 вікі

0.51

 three

0.50

Activations Density 0.819%

references to court cases and legal terms

Code, numbers, or abbreviations after periods or hyphens

multi-language technical terms

No Comments

No Known Activations

references to court cases and legal terms

Code, numbers, or abbreviations after periods or hyphens

multi-language technical terms

No Comments

No Known Activations