INDEX

Explanations

phrases indicating common practices or procedures in various contexts

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

actions followed by purpose

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

httphttps

-0.50

</tfoot>

-0.42

Географиясе

-0.42

-0.41

évaluateur

-0.40

ougou

-0.39

Попис

-0.38

urio

-0.37

Бахар

-0.37

ERICA

-0.37

POSITIVE LOGITS

 becauſe

0.47

 CreateTagHelper

0.45

เพื่อ

0.44

IndentedString

0.43

 հղումներ

0.43

 barbati

0.43

 prefer

0.42

 Kró

0.42

ſy

0.42

 préf

0.41

Activations Density 0.203%

phrases indicating common practices or procedures in various contexts

actions followed by purpose

No Comments

No Known Activations