INDEX

Explanations

connections related to expressions of various states or conditions

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

forms of 'to be' followed by descriptions

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

 kasarigan

-0.42

FormTagHelper

-0.35

 majeurs

-0.34

 mapStateToProps

-0.31

 impuesto

-0.30

Literatuur

-0.30

 pédagogique

-0.29

subpackage

-0.29

AnchorTagHelper

-0.29

 hunne

-0.28

POSITIVE LOGITS

 also

0.62

 always

0.56

 فريبيس

0.49

뀜

0.48

queryInterface

0.47

ValueStyle

0.47

raisals

0.47

 mixed

0.46

 desc

0.46

+#+#

0.46

Activations Density 0.087%

connections related to expressions of various states or conditions

forms of 'to be' followed by descriptions

No Comments

No Known Activations