INDEX

Explanations

specific programming terms and symbols associated with data structure definitions or declarations

oai_token-act-pair · gpt-4o-mini Triggered by @bot

code, hex, type

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

uxxxx

-0.38

 مرئيه

-0.36

 تضيفلها

-0.36

 Unfall

-0.31

 incomplète

-0.31

permanent

-0.31

 braços

-0.31

是多少

-0.30

abstractmethod

-0.30

pão

-0.30

POSITIVE LOGITS

 noDo

0.64

setVerticalGroup

0.56

:✨

0.55

 ivelany

0.54

 cherchés

0.54

 Taktlose

0.52

IVEREF

0.52

principalTable

0.51

 ujednoznacz

0.51

RefNanny

0.51

Activations Density 0.163%

specific programming terms and symbols associated with data structure definitions or declarations

code, hex, type

No Comments

No Known Activations