INDEX

Explanations

references to letters or written communications

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

"letter" followed by a preposition

np_acts-logits-general · gemini-2.0-flash

letter and related communications

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 24-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_24/width_16k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.24.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

 base

-0.29

↵↵

-0.28

cout

-0.27

-0.26

 some

-0.26

 hjel

-0.25

ra

-0.25

 However

-0.25

POSITIVE LOGITS

 letter

1.07

 lettera

1.00

 LETTER

0.94

Letter

0.90

parsedMessage

0.90

 Letter

0.89

 Geiſt

0.86

 letters

0.85

letter

0.84

 initComponents

0.84

Activations Density 0.660%

references to letters or written communications

"letter" followed by a preposition

letter and related communications

No Comments

No Known Activations