INDEX

Explanations

specific scientific categories and experimental design elements

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

labels followed by identifiers

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

-0.30

 otherwise

-0.29

toByteArray

-0.28

SNAPSHOT

-0.26

 يتيمه

-0.26

wyczaj

-0.26

Universal

-0.26

 nueces

-0.25

 inspiradoras

-0.25

 Lordships

-0.25

POSITIVE LOGITS

msgTypes

0.74

tonode

0.71

MigrationBuilder

0.69

RegressionTest

0.64

 thứ

0.64

aarrggbb

0.62

 groupBox

0.62

$_(

0.62

SwingConstants

0.60

findpost

0.60

Activations Density 0.306%

specific scientific categories and experimental design elements

labels followed by identifiers

No Comments

No Known Activations