INDEX

Explanations

formatting related to timestamps and user identifiers

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

JSON key-value pairs

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

帖最后由

-0.60

tvguidetime

-0.54

exitRule

-0.52

ThroughAttribute

-0.52

Derbyniad

-0.50

transQ

-0.50

 sahi

-0.49

الحياه

-0.49

قایناق‌لار

-0.49

AddTagHelper

-0.47

POSITIVE LOGITS

WaitGroup

0.36

clientY

0.35

__':

0.35

ContentAsync

0.34

">—

0.34

</tfoot>

0.33

unicaciones

0.33

жнему

0.32

Архівовано

0.31

urance

0.30

Activations Density 0.003%

formatting related to timestamps and user identifiers

JSON key-value pairs

No Comments

No Known Activations