INDEX

Explanations

code-related constructs and elements, particularly within programming and documentation contexts

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

method name or disclosed

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

 betweenstory

-0.85

 useAppContext

-0.68

.*")]

-0.68

 Generales

-0.59

LookAnd

-0.58

 Wikimedijinoj

-0.58

 שוליים

-0.57

GEBURTS

-0.56

 eşofman

-0.55

AutoField

-0.54

POSITIVE LOGITS

Autoritní

0.50

CppMethod

0.43

tslib

0.42

ngrx

0.39

 useCallback

0.39

ValueStyle

0.38

WEBPACK

0.38

{?}

0.37

 nahilalakip

0.37

 &___

0.36

Activations Density 0.005%

code-related constructs and elements, particularly within programming and documentation contexts

method name or disclosed

No Comments

No Known Activations