INDEX

Explanations

phrases indicating future events or actions

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

owners say, will not play

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

GTCX

-0.60

Personensuche

-0.50

 Савезне

-0.48

:✨

-0.47

oredCriteria

-0.46

ximately

-0.45

fitrión

-0.44

CreateMap

-0.44

 utafitiHapana

-0.43

 pinulongan

-0.42

POSITIVE LOGITS

isContained

0.42

RegistryLite

0.41

 архивлан

0.40

UpInside

0.35

tagHelperRunner

0.35

 entanto

0.34

ToScroll

0.34

MessageTagHelper

0.33

gsSP

0.33

 actionPerformed

0.33

Activations Density 0.006%

phrases indicating future events or actions

owners say, will not play

No Comments

No Known Activations