INDEX

Explanations

words indicating variability or instances of exceptions in situations

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

some ... and others

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

LEncoder

-0.48

 pinulongan

-0.47

 viewDidLoad

-0.43

mbggenerated

-0.42

 recep

-0.42

 INFORM

-0.40

 transfieras

-0.39

 autorytatywna

-0.39

mybatisplus

-0.39

 surla

-0.38

POSITIVE LOGITS

 سكانية

0.52

 juist

0.47

 Lordships

0.45

 للاسماء

0.40

IVEREF

0.40

 sebaliknya

0.40

 deles

0.40

 unfortunately

0.39

➋

0.38

localctx

0.38

Activations Density 0.022%

words indicating variability or instances of exceptions in situations

some ... and others

No Comments

No Known Activations