INDEX

Explanations

patterns related to reporting and statements in discussions

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

Follows punctuation or articles (the, a, an)

np_acts-logits-general · gemini-2.0-flash

end of sentences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.18.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

 myſelf

-1.01

 Monfieur

-0.96

Efq

-0.95

 Houſe

-0.85

 Theſe

-0.85

 whoſe

-0.85

 ſhe

-0.82

 Reſ

-0.82

 ſeveral

-0.80

^(@)

-0.79

POSITIVE LOGITS

 authors

0.74

 according

0.64

the

0.63

文中

0.63

 report

0.62

According

0.58

 According

0.58

 noted

0.58

Bronnen

0.58

 stated

0.56

Activations Density 0.234%

patterns related to reporting and statements in discussions

Follows punctuation or articles (the, a, an)

end of sentences

No Comments

No Known Activations