INDEX

Explanations

personal anecdotes and opinions, signaled by words like "I", "you", "we", "want," "love," "would," and "hear."

oai_token-act-pair · gemini-2.0-flash

here

np_max-act-logits · gemini-2.0-flash

New Auto-Interp

Configuration

google/gemma-scope-2b-pt-transcoders/layer_25/width_16k/average_l0_41

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.25.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

enumii

-0.76

 heard

-0.73

Heard

-0.71

 Heard

-0.70

heard

-0.66

TestingModule

-0.59

 Hearing

-0.56

RunWith

-0.54

Hearing

-0.52

 hearing

-0.51

POSITIVE LOGITS

 here

3.34

here

2.69

 Here

2.56

 aquí

2.47

Here

2.47

ici

2.27

 HERE

2.25

 هنا

2.20

 здесь

2.19

 aqui

2.14

Activations Density 4.261%

personal anecdotes and opinions, signaled by words like "I", "you", "we", "want," "love," "would," and "hear."

here

No Comments

No Known Activations