INDEX

Explanations

articles, possessives and references to multiple things being counted

oai_token-act-pair · gemini-2.0-flash

two or three

np_max-act-logits · gemini-2.0-flash

New Auto-Interp

Configuration

google/gemma-scope-2b-pt-transcoders/layer_20/width_16k/average_l0_11

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.20.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



-0.77

 resourceCulture

-0.77

 preuves

-0.75

Efq

-0.73

Hochspringen

-0.73

 Theſe

-0.72

 verksamhet

-0.69

 étoient

-0.67

 dieux

-0.63

 cérami

-0.62

POSITIVE LOGITS

or

0.54

یش

0.50

0.49

 distinct

0.48

رشف

0.48

vero

0.48

of

0.47

TagMode

0.47

KELEY

0.47

legged

0.46

Activations Density 4.983%

articles, possessives and references to multiple things being counted

two or three

No Comments

No Known Activations