INDEX

Explanations

medical and health-related terminology

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

actions and concepts

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

 Chwiliwch

-0.51

AndEndTag

-0.47

 Paglinawan

-0.47

CppMethod

-0.43

انجليز

-0.42

MessageOf

-0.42

 Infórmanos

-0.42

igshid

-0.41

 чему

-0.41

 للمعارف

-0.40

POSITIVE LOGITS

 Logement

0.47

DeleteBehavior

0.41

 Roskov

0.40

ViewFeatures

0.38

ViewInit

0.36

DataBinding

0.36

 prosím

0.35

 tampoco

0.35

expandindo

0.35

rawDesc

0.35

Activations Density 0.337%

medical and health-related terminology

actions and concepts

No Comments

No Known Activations