INDEX

Explanations

code-related terms or concepts

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

LaTeX math mode and commands

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

 Infórmanos

-0.44

 ویکی‌پدیای

-0.36

 <<<<<<<<<<<<<<

-0.36

miyor

-0.35

displayquote

-0.35

########.

-0.34

 فريبيس

-0.33

 Pembangunan

-0.33

ieť

-0.32

 spie

-0.32

POSITIVE LOGITS

Abit

0.56

Voci

0.54

principalColumn

0.54

 ModelExpression

0.52

 okuyayım

0.52

তথ্যসূত্র

0.51

ParallelGroup

0.51

 kasarigan

0.49

SourceChecksum

0.49

migrationBuilder

0.48

Activations Density 0.095%

code-related terms or concepts

LaTeX math mode and commands

No Comments

No Known Activations