INDEX

Explanations

references to the word "one" and its variations in context

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

specific ones or one

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

Portály

-0.53

-------

-0.47

 überhaupt

-0.43

 wikipagina

-0.41

 inderdaad

-0.40

 صرف

-0.39

lysis

-0.39

 Storey

-0.39

 Miß

-0.38

POSITIVE LOGITS

 ones

0.64

ThroughAttribute

0.64

Ones

0.60

 Ones

0.59

 Celui

0.57

 հղումներ

0.55

 whose

0.54

 الرياضيه

0.54

Билгалдахарш

0.53

 Ceux

0.53

Activations Density 0.012%

references to the word "one" and its variations in context

specific ones or one

No Comments

No Known Activations