INDEX

Explanations

affirmations or expressions of positivity

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 5-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_5/width_16k/average_l0_68

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.5.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 multer

-0.68

 Cura

-0.67

multer

-0.67

<tr>

-0.66

 pilar

-0.65

#+#

-0.65

 Chatham

-0.65

 Limburg

-0.64

 Brim

-0.64

ZE

-0.64

POSITIVE LOGITS

 actually

1.84

actually

1.67

Actually

1.65

 Actually

1.62

 ACTUALLY

1.40

 faktiskt

1.08

 faktisk

1.07

 sebenarnya

1.02

 fact

0.96

 tatsächlich

0.95

Activations Density 0.077%

affirmations or expressions of positivity

No Comments

No Known Activations