INDEX

Explanations

phrases related to evaluating conditions and outcomes in research or studies

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

user turn markers

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

AnimationsModule

-0.46

aksikan

-0.37

<eos>

-0.32

InputTagHelper

-0.32

 CommonModule

-0.30

ategy

-0.30

 itemprop

-0.30

 εκεί

-0.30

äksi

-0.29

 sienta

-0.29

POSITIVE LOGITS

ftagPool

0.71

 initComponents

0.69

RTLR

0.67

tagPool

0.65

Obrázky

0.61

Билгалдахарш

0.58

Referanser

0.57

HtmlAttribute

0.57

Chham

0.57

kháu

0.56

Activations Density 0.052%

phrases related to evaluating conditions and outcomes in research or studies

user turn markers

No Comments

No Known Activations