INDEX

Explanations

specific time periods and dates within historical contexts

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

years in ranges

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

permum

-0.63

parency

-0.50

曖昧さ回避

-0.48

iscope

-0.47

weck

-0.46

ectoria

-0.46

vician

-0.46

rotter

-0.44

StoryboardSegue

-0.44

ogy

-0.43

POSITIVE LOGITS

 Infórmanos

0.48

Abitanti

0.43

endaft

0.38

AnchorStyles

0.38

audiovisuel

0.35

例句

0.34

AndEndTag

0.34

 Roskov

0.34

 supuestamente

0.34

EndContext

0.34

Activations Density 0.026%

specific time periods and dates within historical contexts

years in ranges

No Comments

No Known Activations