INDEX

Explanations

terms related to media playback and official declarations or announcements

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

proclamations and playback

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

 aéro

-0.46

 fire

-0.45

ỳnh

-0.41

 rêver

-0.40

 moldes

-0.38

กรณ์

-0.38

 meurt

-0.38

gun

-0.37

hütte

-0.37

 rigs

-0.37

POSITIVE LOGITS

textContent

1.43

Personendaten

0.67

 MacLeod

0.64

ugier

0.63

OLEAN

0.62

cyon

0.61

textLabel

0.60

lcccccc

0.60

ongyang

0.59

existsSync

0.59

Activations Density 0.009%

terms related to media playback and official declarations or announcements

proclamations and playback

No Comments

No Known Activations