INDEX

Explanations

references to technical or procedural terminology

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

suffering from | technically difficult

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

脚注の使い方

-0.55

 مرئيه

-0.50

GraphicsUnit

-0.50

extAlignment

-0.44

Mel

-0.42

enteri

-0.40

bri

-0.40

 escape

-0.39

providedIn

-0.39

TSR

-0.39

POSITIVE LOGITS

WebElementEntity

0.52

 Universitaria

0.48

 stavu

0.48

 cenar

0.48

 skolan

0.48

AsUp

0.46

 varandra

0.46

 paille

0.46

 campamento

0.46

 initComponents

0.45

Activations Density 0.143%

references to technical or procedural terminology

suffering from | technically difficult

No Comments

No Known Activations