INDEX

Explanations

phrases related to quality

oai_token-act-pair · gemini-2.0-flash

condition

np_max-act-logits · gemini-2.0-flash

New Auto-Interp

Configuration

google/gemma-scope-2b-pt-transcoders/layer_22/width_16k/average_l0_15

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.22.ln2.hook_normalized

Architecture

jumprelu_transcoder

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 étoit

-0.62

in

-0.60

 oignons

-0.57

 enfans

-0.57

 asiatique

-0.56

 avoient

-0.56

 auroit

-0.54

 jeunesse

-0.54

 debole

-0.54

 répondit

-0.52

POSITIVE LOGITS

 oprot

0.62

पया

0.61

 shape

0.59

 iNdEx

0.57

@"/

0.57

PyExc

0.57

addGap

0.57

bukkit

0.56

Datuak

0.56

Personensuche

0.55

Activations Density 0.891%

phrases related to quality

condition

No Comments

No Known Activations