INDEX

Explanations

phrases related to completion or termination of activities or states

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

when finished or ready

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

 uitge

-0.33

 including

-0.31

masuk

-0.30

 Masuk

-0.30

 unique

-0.30

 incluindo

-0.30

 couvre

-0.30

 yük

-0.29

EEP

-0.29

exp

-0.29

POSITIVE LOGITS

 الحره

0.67

expandindo

0.60

 للاسماء

0.54

parsedMessage

0.53

verwijspagina

0.53

tagHelperRunner

0.53

RegressionTest

0.51

CodeAttribute

0.51

oa̍t

0.50

 aDecoder

0.50

Activations Density 0.018%

phrases related to completion or termination of activities or states

when finished or ready

No Comments

No Known Activations