INDEX

Explanations

references to parliamentary systems and terms

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

parliament

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

cured

-0.43

 Kiwi

-0.43

Dear

-0.41

No

-0.41

sore

-0.41

drip

-0.41

 Lazy

-0.41

creepy

-0.40

Seen

-0.40

sli

-0.40

POSITIVE LOGITS

 Parliament

1.13

 parliament

1.08

Parliament

0.98

 Parlament

0.90

 parli

0.89

 Parli

0.87

parliament

0.83

 parlament

0.82

liament

0.77

 Parlement

0.75

Activations Density 0.006%

references to parliamentary systems and terms

parliament

No Comments

No Known Activations