INDEX

Explanations

contractions and negations in the text

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

United States

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

fully

-0.32

ery

-0.27

 obicei

-0.27

 نظ

-0.27

ig

-0.26

ph

-0.26

ary

-0.25

gesetzt

-0.24

 precum

-0.24

POSITIVE LOGITS

 الحره

0.73

tvguidetime

0.70

KommentareTeilen

0.69

 Wiktionnaire

0.67

MigrationBuilder

0.66

CppMethod

0.65

 Administrativna

0.65

省市镇

0.65



0.64

 Meksiku

0.63

Activations Density 0.023%

contractions and negations in the text

United States

No Comments

No Known Activations