INDEX

Explanations

instances of the phrase "used to" as a marker for past habits or experiences

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

past states or habits

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

 pleaſure

-0.56

 itſelf

-0.54

 lumineuse

-0.53

 intérieure

-0.53

 investigación

-0.51

 Entsch

-0.50

 chrétienne

-0.48

hæng

-0.48

 investigação

-0.47

 hjel

-0.47

POSITIVE LOGITS

be

0.56

 Rusty

0.55

 rusty

0.55

staticmethod

0.54

cesz

0.53

oldValue

0.53

جوايز

0.53

 Administrativna

0.52

 rust

0.52

 costuma

0.51

Activations Density 0.022%

instances of the phrase "used to" as a marker for past habits or experiences

past states or habits

No Comments

No Known Activations