INDEX

Explanations

references to prior knowledge or information and also indications of existing conditions or precedents

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

already existing

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

 Nachteile

-0.34

 interesan

-0.33

 pohon

-0.32

 adopción

-0.31

 sillas

-0.30

 vrst

-0.30

 democrá

-0.29

 lámina

-0.29

Caracteristici

-0.29

Przypisy

-0.28

POSITIVE LOGITS

 typelib

0.62

WithIOException

0.62

 محفوظة

0.59

bkz

0.56

migrationBuilder

0.56

 OkHttpClient

0.56

addCriterion

0.56

 Codable

0.55

EndInit

0.55

webElementXpaths

0.54

Activations Density 0.046%

references to prior knowledge or information and also indications of existing conditions or precedents

already existing

No Comments

No Known Activations