INDEX

Explanations

code-related keywords and structures within programming languages

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

code definitions and syntax

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

 sauvages

-0.35

 cielos

-0.33

 contigo

-0.32

 corazones

-0.32

 povezave

-0.31

 poveznice

-0.30

Specyfikacja

-0.30

 lainnya

-0.30

 claros

-0.29

样

-0.28

POSITIVE LOGITS

 ब्रेकडाउन

0.65

ⓧ

0.63

новништво

0.60

 дописавши

0.59

farande

0.59

帖最后由

0.58

 otomatig

0.57

OGND

0.54

Personensuche

0.54

 dagger

0.52

Activations Density 0.070%

code-related keywords and structures within programming languages

code definitions and syntax

No Comments

No Known Activations