INDEX

Explanations

numerical values or identifiers related to data or processes

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

hexadecimal codes and kernel

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

isches

-0.38

 campaigns

-0.38

 disruptions

-0.37

 defences

-0.36

eventType

-0.35

 purpose

-0.34

 newspapers

-0.34

 mental

-0.34

 friction

-0.34

 Kong

-0.33

POSITIVE LOGITS

setVerticalGroup

0.73

évaluateur

0.71

 mijne

0.67

Datuak

0.66

KURZBESCHREIBUNG

0.63

Personendaten

0.61

ManagedBean

0.60

 autorytatywna

0.59

httphttps

0.59

 مشين

0.59

Activations Density 0.001%

numerical values or identifiers related to data or processes

hexadecimal codes and kernel

No Comments

No Known Activations