INDEX

Explanations

extreme adjectives and their intensifiers

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

certainty or completeness

np_acts-logits-general · gemini-2.0-flash

intensifiers followed by descriptors

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

 الرياضيه

-0.55

 Inscrivez

-0.54

Билгалдахарш

-0.46

 sindic

-0.46

Altri

-0.45

 gruesa

-0.45

Память

-0.45

 estekak

-0.45

SerializeField

-0.45

paravant

-0.44

POSITIVE LOGITS

 completely

0.96

 totally

0.96

 absolutely

0.93

totally

0.87

 utterly

0.85

 absolute

0.84

Completely

0.83

 Completely

0.82

completely

0.82

Absolutely

0.80

Activations Density 0.161%

extreme adjectives and their intensifiers

certainty or completeness

intensifiers followed by descriptors

No Comments

No Known Activations