INDEX

Explanations

punctuation and formatting symbols indicating structure within the text

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

news attribution

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

 central

-0.35

inheritDoc

-0.33

 Enix

-0.33

 automatically

-0.31

 Nederland

-0.30

ceğ

-0.30

iola

-0.30

 consistently

-0.29

 adequately

-0.29

 endlessly

-0.29

POSITIVE LOGITS

 arşivlendi

0.73



0.70

 AssemblyProduct

0.68

Kjelder

0.63

GEBURTSDATUM

0.60

KommentareTeilen

0.60

Tikang

0.59

RegressionTest

0.58

 EconPapers

0.57

 disambiguazione

0.56

Activations Density 0.001%

punctuation and formatting symbols indicating structure within the text

news attribution

No Comments

No Known Activations