INDEX

Explanations

is able to detect the presence of the word "error" (エラー) as well as some Japanese sentence structure related to encountering errors

oai_token-act-pair · gemini-2.0-flash

programming and 日本語

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

出版年

-0.63

wahati

-0.62

Бахар

-0.58

ształ

-0.57

 '\\;'

-0.56

 useEffect

-0.56

marginHorizontal

-0.55

 Great

-0.54

ometrial

-0.54

uarga

-0.53

POSITIVE LOGITS

ers

0.55

 Wikimedijinoj

0.55

featureID

0.54

ViewFeatures

0.52

ings

0.44

ENDIAN

0.44

CodeAnalysis

0.44

WriteAttribute

0.43

Tikang

0.43

Източници

0.41

Activations Density 0.387%

is able to detect the presence of the word "error" (エラー) as well as some Japanese sentence structure related to encountering errors

programming and 日本語

No Comments

No Known Activations

is able to detect the presence of the word "error" (エラー) as well as some Japanese sentence structure related to encountering errors

programming and 日本語

No Comments

No Known Activations