INDEX

Explanations

references to judicial opinions and court-related terms

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Py

-0.45

 rowspan

-0.44

-0.43

ok

-0.39

by

-0.39

か

-0.39

 psycopg

-0.39

 trick

-0.38

ffindor

-0.38

rofoam

-0.38

POSITIVE LOGITS

uxxxx

0.99

Personendaten

0.95

 disambiguazione

0.91

 surla

0.91

GEBURTSDATUM

0.85

 autorytatywna

0.83

WriteTagHelper

0.79

complexContent

0.77

 AssemblyCulture

0.77

BibitemShut

0.73

Activations Density 0.001%

references to judicial opinions and court-related terms

No Comments

No Known Activations