INDEX

Explanations

abbreviations and acronyms related to organizations and medical conditions

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

acronyms in parentheses

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

Datuak

-0.74

 يتيمه

-0.73

 PeEnEo

-0.69

 дописавши

-0.69

 صوتيه

-0.66

 Roskov

-0.65

 pinulongan

-0.65

IUrlHelper

-0.63

 nonUne

-0.60

rawDesc

-0.59

POSITIVE LOGITS

 relógio

0.40

 parachoque

0.34

 HttpNotFound

0.32

 capuz

0.32

客様

0.31

 fraî

0.31

 quidem

0.31

Geschiedenis

0.31

 sujets

0.31

 prêtres

0.30

Activations Density 0.039%

abbreviations and acronyms related to organizations and medical conditions

acronyms in parentheses

No Comments

No Known Activations