INDEX

Explanations

phrases indicating research methodology or study intentions

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

reporting research findings

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

 дописавши

-0.32

 invokingState

-0.32

 molho

-0.32

 løpet

-0.30

 consideración

-0.30

 DialogInterface

-0.30

Garantía

-0.29

ฆ

-0.28

 döv

-0.28

WebElementEntity

-0.28

POSITIVE LOGITS

transQ

0.61

OGND

0.59

Aims

0.58

 Aims

0.54

 research

0.54

 ricerca

0.52

 recherche

0.52

RefNanny

0.52

setVerticalGroup

0.51

stateMutability

0.51

Activations Density 0.249%

phrases indicating research methodology or study intentions

reporting research findings

No Comments

No Known Activations