INDEX

Explanations

references to authors and their credentials in academic contexts

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

Initial uppercase letters

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

rungsseite

-0.57

 surla

-0.55

PyExc

-0.50

 Waray

-0.49

 ویکی‌پدی

-0.47

ftagPool

-0.46

 للاسماء

-0.46

曖昧さ回避

-0.45

ngrx

-0.43

omitempty

-0.42

POSITIVE LOGITS

Sac

0.48

tagHelperRunner

0.47

Gr

0.45

Hab

0.43

 Office

0.43

Har

0.42

Ant

0.42

ữ

0.42

Fo

0.41

Ban

0.41

Activations Density 0.043%

references to authors and their credentials in academic contexts

Initial uppercase letters

No Comments

No Known Activations