INDEX

Explanations

phrases indicating medical evaluations and physiological conditions

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

preceding technical terms or names

np_acts-logits-general · gemini-2.0-flash

significant, difficulty, hurt, verified, most

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.18.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

 sufficiently

-0.88

 substantially

-0.79

 briefly

-0.79

 diligently

-0.78

 principally

-0.75

 slightly

-0.75

 primarily

-0.75

 abbastanza

-0.74

 significantly

-0.74

 essentiellement

-0.73

POSITIVE LOGITS

saraba

0.65

UnusedPrivate

0.55

(>

0.55

 PyTuple

0.50

LElement

0.49

 sekali

0.46

WebServlet

0.45

(~

0.45

 насељу

0.44

 Infatti

0.44

Activations Density 0.888%

phrases indicating medical evaluations and physiological conditions

preceding technical terms or names

significant, difficulty, hurt, verified, most

No Comments

No Known Activations