INDEX

Explanations

tokens or symbols typically used in mathematical expressions or programming contexts

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

user turn start

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

ConstraintMaker

-0.65

 Taktlose

-0.57

Vidite

-0.57

<bos>

-0.49

XmlAccessorType

-0.46

पया

-0.42

 препратки

-0.42

 SqlDbType

-0.40

$_['

-0.40

TestTools

-0.40

POSITIVE LOGITS

IndentedString

0.43

HomeAsUpEnabled

0.41

fjspx

0.41

CppMethod

0.40

過

0.38

tagext

0.37

bitField

0.35

 Placeholder

0.35

lapsingToolbar

0.35

Initializable

0.35

Activations Density 0.000%

tokens or symbols typically used in mathematical expressions or programming contexts

user turn start

No Comments

No Known Activations