INDEX

Explanations

words associated with visual elements or imaging

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

in languages other than English

np_acts-logits-general · gemini-2.0-flash

Indonesian/Malay phrases and English terms

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 24-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_24/width_16k/average_l0_73

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.24.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

 verf

-0.48

 Entire

-0.47

entire

-0.47

]-->

-0.45

+#+

-0.45

herjee

-0.45

</caption>

-0.44

Ainsi

-0.44

GEBURTS

-0.43

 entire

-0.42

POSITIVE LOGITS

Personendaten

0.71

AccessorTable

0.66

InstrumentedTest

0.61

ConstraintMaker

0.60

 inget

0.59

 Biôgrafia

0.57

ftagPool

0.57

 protoimpl

0.55

PyTuple

0.55

 kaarangay

0.54

Activations Density 0.030%

words associated with visual elements or imaging

in languages other than English

Indonesian/Malay phrases and English terms

No Known Activations