INDEX

Explanations

keywords related to programming functions and database queries

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

criminal cases

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

Popular

-0.41

AutoField

-0.40

<h1>

-0.36

“…

-0.36

<h2>

-0.35

 pueda

-0.35

"...

-0.35

Allgemeine

-0.34

<strong>

-0.34

POSITIVE LOGITS

TagMode

0.70

 AssemblyCulture

0.59

 كومونز

0.59

 MainAxisSize

0.59

exitRule

0.56

KommentareTeilen

0.56

出版年

0.56

principalColumn

0.54

IntoConstraints

0.54

脚注の使い方

0.54

Activations Density 0.054%

keywords related to programming functions and database queries

criminal cases

No Comments

No Known Activations