INDEX

Explanations

negations and qualifiers indicating uncertainty or exceptions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 33-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_33/width_16k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.33.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

 виправивши

-0.39

}*/

-0.33

 Pembangunan

-0.32

})*/

-0.32

otherwise

-0.32

 finalement

-0.31

 otherwise

-0.31

 Lingkungan

-0.31

queryInterface

-0.31

whatever

-0.30

POSITIVE LOGITS

 reason

0.86

 dearth

0.77

 danger

0.73

 shortage

0.73

 possibility

0.73

 tendency

0.71

 inherent

0.67

possibility

0.67

 discrepancy

0.67

 difference

0.67

Activations Density 0.325%

negations and qualifiers indicating uncertainty or exceptions

No Comments

No Known Activations