INDEX

Explanations

specific names and terms related to substances, places, and categories

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

specific entities and places

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

Personendaten

-0.61

 Audiodateien

-0.59

Datuak

-0.54

 nonUne

-0.54

OGND

-0.52

Bibliograf

-0.48

oa̍t

-0.48

 arşivlendi

-0.47

ArrowToggle

-0.47

Lähteet

-0.46

POSITIVE LOGITS

irtschaft

0.39

addGap

0.38

usitis

0.36

 Favre

0.35

 respectively

0.34

这两个

0.33

Ens

0.33

 губ

0.33

 eset

0.33

 despacho

0.33

Activations Density 0.052%

specific names and terms related to substances, places, and categories

specific entities and places

No Comments

No Known Activations