INDEX

Explanations

references to past experiences and actions related to reading or using products

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

research actions performed

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

出版年

-0.75

 surla

-0.60

 plufieurs

-0.58

 dezelve

-0.58

parsedMessage

-0.57

 informée

-0.56

 itſelf

-0.56

 krwi

-0.54

 queſta

-0.54

Personensuche

-0.52

POSITIVE LOGITS

Sam

0.47

sam

0.46

bes

0.44

Sam

0.42

SequentialGroup

0.41

roid

0.40

pin

0.39

 opinion

0.39

 Kling

0.39

vim

0.38

Activations Density 0.083%

references to past experiences and actions related to reading or using products

research actions performed

No Comments

No Known Activations