INDEX

Explanations

numerical data and references to experimental results or counts

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

Greek letters and numbers

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

seamnă

-0.53

Lähteet

-0.50

 mijne

-0.50

pecabe

-0.48

afficheront

-0.48

PMailer

-0.47

KURZBESCHREIBUNG

-0.47

 îna

-0.47

 zijne

-0.47

 вікісторінку

-0.46

POSITIVE LOGITS

 defaultstate

0.48

fromnode

0.47

 intermediate

0.42

 intermediary

0.38

providedIn

0.38

 intermedi

0.35

 protest

0.34

Green

0.31

__);

0.31

Intermediate

0.31

Activations Density 0.130%

numerical data and references to experimental results or counts

Greek letters and numbers

No Comments

No Known Activations