INDEX

Explanations

phrases related to conditions and outcomes in the context of medical or procedural discussions

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

function and possible changes

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

 للاسماء

-0.37

Lundi

-0.33

 betweenstory

-0.29

 âmes

-0.29

LayoutStyle

-0.29

ajinan

-0.29

 horario

-0.28

 horarios

-0.28

 الدولى

-0.28

 flexível

-0.27

POSITIVE LOGITS

mybatisplus

0.70

RegressionTest

0.59

Enllaces

0.59

ьаж

0.58

StoreMessageInfo

0.58

 snippetHide

0.56

पया

0.50

ViewImports

0.49

nodoc

0.48

JspWriter

0.48

Activations Density 0.051%

phrases related to conditions and outcomes in the context of medical or procedural discussions

function and possible changes

No Comments

No Known Activations