INDEX

Explanations

parentheses and their usage in code-like structures

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

function parameter lists or definitions

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

 overras

-0.56

 julk

-0.43

 poème

-0.41

 épu

-0.40

 vété

-0.40

 noodzake

-0.39

private

-0.38

 refusé

-0.37

elemField

-0.37

 kræ

-0.37

POSITIVE LOGITS

脚注の使い方

0.60

ruptedException

0.55

 kasarigan

0.55

 banjo

0.54

BindingResult

0.52

 المرجع

0.51

béco

0.50

 EventArgs

0.50

gamon

0.50

 binder

0.49

Activations Density 0.008%

parentheses and their usage in code-like structures

function parameter lists or definitions

No Comments

No Known Activations