INDEX

Explanations

structured or formatted data, possibly related to programming or software documentation

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

code delimiters and separators

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

 formel

-0.32

mes

-0.32

 Exactos

-0.32

apl

-0.31

ม่

-0.31

今年も

-0.30

 NgModule

-0.30

 amarillas

-0.30

StyleSheet

-0.30

@[+][

-0.30

POSITIVE LOGITS

DoubleQuotes

0.62

DockStyle

0.58

VersionUID

0.57

HtmlAttribute

0.56

 AttributeSet

0.56

yntaxException

0.56

 <>",

0.55

RefNanny

0.54

complexContent

0.53

principalTable

0.52

Activations Density 0.074%

structured or formatted data, possibly related to programming or software documentation

code delimiters and separators

No Comments

No Known Activations