INDEX

Explanations

negations and contractions in the text

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

numbers in text

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_65k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

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

المناصب

-0.76

 resourceCulture

-0.74

 ExecuteAsync

-0.73

OrNil

-0.70

tvguidetime

-0.68

 EconPapers

-0.67

 defaultstate

-0.67

GEBURTSDATUM

-0.66

Hauptartikel

-0.64

ArrowToggle

-0.64

POSITIVE LOGITS

 this

0.62

it

0.52

 another

0.51

 program

0.51

 argument

0.50

 what

0.50

addView

0.50

0.49

 report

0.49

 quel

0.47

Activations Density 0.057%

negations and contractions in the text

numbers in text

No Comments

No Known Activations