INDEX

Explanations

dimensions and sizes often related to graphical representations or mathematical descriptions

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

Appearing before numbers/coordinates

np_acts-logits-general · gemini-2.0-flash

dimensions like n x n

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

expandindo

-0.79

exitRule

-0.65

 ویکی‌پدی

-0.63

RegressionTest

-0.61

 zoude

-0.60

%)$

-0.59

 Houſe

-0.59

 intStringLen

-0.58

transQ

-0.58

 Anſ

-0.57

POSITIVE LOGITS

RenderAtEndOf

0.34

 acompañado

0.33

 bezpieczeństwa

0.33

 ligiloj

0.33

NegativeButton

0.33

 parceria

0.33

 razão

0.32

 leitura

0.32

 suspendu

0.31

 obok

0.30

Activations Density 0.510%

dimensions and sizes often related to graphical representations or mathematical descriptions

Appearing before numbers/coordinates

dimensions like n x n

No Comments

No Known Activations