INDEX

Explanations

elements related to instructional steps or procedural guidelines

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 34-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_34/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.34.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

 also

-0.85

 tiež

-0.75

 likewise

-0.73

also

-0.72

 ayrıca

-0.70

 також

-0.67

 myös

-0.66

 också

-0.64

 other

-0.63

 также

-0.63

POSITIVE LOGITS

 Firstly

0.81

Firstly

0.80

 ſta

0.75

 Jefus

0.74

 للاسماء

0.71

 cauſe

0.70

 ſtate

0.67

 Verſ

0.66

ſelf

0.66

 ſche

0.66

Activations Density 0.743%

elements related to instructional steps or procedural guidelines

No Comments

No Known Activations