INDEX

Explanations

keywords and phrases related to categories, classifications, and institutional frameworks

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

formats and categories

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

évaluateur

-0.42

 diting

-0.37

 موا

-0.37

şört

-0.36

 adventure

-0.35

 byte

-0.35

 rost

-0.34

ならない

-0.34

 restoration

-0.34

 Packers

-0.34

POSITIVE LOGITS

AndEndTag

0.56

 通販

0.52

tagHelperRunner

0.52

IUrlHelper

0.46

цездатний

0.45

SBATCH

0.44

艋

0.43

__':

0.42

astéroïdes

0.42

principalTable

0.41

Activations Density 0.017%

keywords and phrases related to categories, classifications, and institutional frameworks

formats and categories

No Comments

No Known Activations