INDEX

Explanations

references to letters and formal written communication

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

"letter" or its variations

np_acts-logits-general · gemini-2.0-flash

letter

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

 normalidad

-0.34

 Bhopal

-0.31

Collapse

-0.31

 disposer

-0.30

 tâche

-0.30

 gärna

-0.30

 mestre

-0.30

 Domes

-0.30

 humains

-0.30

 differenti

-0.30

POSITIVE LOGITS

 letter

1.21

 letters

1.09

Letter

1.03

 Letter

1.03

 LETTER

0.98

 lettera

0.96

 LETTERS

0.96

 Letters

0.94

Letters

0.94

PMailer

0.90

Activations Density 0.248%

references to letters and formal written communication

"letter" or its variations

letter

No Comments

No Known Activations