INDEX

Explanations

references to archives and archiving processes

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

archive and archives

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

 Nelly

-0.51

 didst

-0.50

 Mussolini

-0.49

uterte

-0.48

odd

-0.48

 StringVar

-0.46

ctron

-0.46

 Nissan

-0.45

 Yeon

-0.45

!(

-0.44

POSITIVE LOGITS

 archive

1.09

 Archive

1.03

Archive

1.02

 archives

0.98

archive

0.98

 archiving

0.90

 Archives

0.88

 arkiv

0.84

 archival

0.83

ARCHIVE

0.82

Activations Density 0.008%

references to archives and archiving processes

archive and archives

No Comments

No Known Activations