INDEX

Explanations

phrases indicating speculation or uncertainty

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

times gone by

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

 TikTok

-0.50

 onboarding

-0.48

 Edmond

-0.46

ofür

-0.45

 https

-0.45

MenuInflater

-0.45

tiktok

-0.44

 Dutton

-0.44

 airpods

-0.43

 Tiktok

-0.43

POSITIVE LOGITS

 damals

0.63

 damal

0.52

RegressionTest

0.51

当時

0.50

当時は

0.50

 kasarigan

0.50

当年

0.49

 damaligen

0.48

rawDesc

0.47

当時の

0.47

Activations Density 0.115%

phrases indicating speculation or uncertainty

times gone by

No Comments

No Known Activations