INDEX

Explanations

numerical data and references to figures and indices

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

image width markdown

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

tagHelperRunner

-0.82

hyrchwyd

-0.57

tangentMode

-0.56

enterOuterAlt

-0.54

rungsseite

-0.54

OuterAlt

-0.52

 nahilalakip

-0.52

WriteTagHelper

-0.49

 autorytatywna

-0.48

WebElementEntity

-0.48

POSITIVE LOGITS

 gleichzeitig

0.40

 ครับ

0.39

 sequência

0.36

Personendaten

0.36

 Jurí

0.36

 simultaneously

0.35

 inneren

0.34

 concurrently

0.34

ณ์

0.34

vidéo

0.33

Activations Density 0.023%

numerical data and references to figures and indices

image width markdown

No Comments

No Known Activations