INDEX

Explanations

phrases related to legal or medical actions and responsibilities

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

when something happens

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

</tfoot>

-0.44

✭✭

-0.42

.*")]

-0.41

ﬅ

-0.40

Supplies

-0.40

 Спи

-0.40

autoreleasepool

-0.39

 shaft

-0.39

bpf

-0.39

 esprits

-0.39

POSITIVE LOGITS

 when

0.64

 wanneer

0.61

 Infór

0.58

 ketika

0.53

when

0.50

 når

0.50

ServletException

0.48

 wenn

0.48

if

0.48

 lorsqu

0.47

Activations Density 0.047%

phrases related to legal or medical actions and responsibilities

when something happens

No Comments

No Known Activations