INDEX

Explanations

phrases and terms related to legal and procedural contexts

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

preceding "with"

np_acts-logits-general · gemini-2.0-flash

relating entities

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

DoubleQuotes

-0.51

 Unterschied

-0.45

Gemeinsame

-0.45

 spokesperson

-0.45

 alene

-0.44

Including

-0.43

ঁ

-0.42

Todos

-0.42

-0.41

per

-0.41

POSITIVE LOGITS

 existing

0.90

 eachother

0.90

 nhau

0.85

 other

0.83

 varandra

0.80

ISupport

0.78

 bestehenden

0.78

 bestehende

0.77

__':

0.77

OGND

0.76

Activations Density 0.729%

phrases and terms related to legal and procedural contexts

preceding "with"

relating entities

No Comments

No Known Activations