INDEX

Explanations

references to high levels or rates of various medical or health-related phenomena

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

high degree, chance, morbidity, correlation, density, opinion, death, cost, rates, surface

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

AnimationsModule

-0.41

плек

-0.41

 someplace

-0.39

Zas

-0.38

actionBar

-0.38

JF

-0.37

 somewhere

-0.36

JL

-0.35

 Winslow

-0.35

 whispers

-0.35

POSITIVE LOGITS

 highly

0.68

 rate

0.64

 high

0.62

principalColumn

0.62

highly

0.60

 Highly

0.56

Highly

0.56

 taux

0.56

 vysok

0.54

 pinulongan

0.54

Activations Density 0.047%

references to high levels or rates of various medical or health-related phenomena

high degree, chance, morbidity, correlation, density, opinion, death, cost, rates, surface

No Comments

No Known Activations