INDEX

Explanations

code or syntax elements related to programming logic and structures

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

solving math equations

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

eader

-0.40

palen

-0.35

 whistled

-0.35

 pane

-0.35

 CreateTagHelper

-0.34

 repug

-0.34

ederen

-0.34

izin

-0.34

son

-0.34

ise

-0.34

POSITIVE LOGITS

 виправивши

0.70

 indígen

0.52

 Majefty

0.50

 Italij

0.49

 poitrine

0.49

 ویکی‌پدی

0.48

 lørdag

0.48

 Infór

0.47

 pouvoit

0.46

 Roskov

0.46

Activations Density 0.003%

code or syntax elements related to programming logic and structures

solving math equations

No Comments

No Known Activations