INDEX

Explanations

references to food-related items and cooking instructions

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

trained CNN

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



-0.34

per

-0.33

 ditem

-0.32

 neler

-0.32

 upstream

-0.32

 Aprend

-0.32

 Mühl

-0.31

明天

-0.31

 Sosten

-0.30

 leading

-0.30

POSITIVE LOGITS

 للاسماء

0.65

StoryboardSegue

0.60

 ujednoznacz

0.56

smithy

0.56

出版年

0.56

 ब्रेकडाउन

0.55

 المعيارى

0.54

 dieſe

0.52

✨:

0.51

 dieſen

0.51

Activations Density 0.047%

references to food-related items and cooking instructions

trained CNN

No Comments

No Known Activations