INDEX

Explanations

the term "Ten" and its variations, suggesting a focus on numerical or measurement-related contexts

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

ciorys

-0.33

Cyfarwyddwr

-0.32

 Bagi

-0.32

scribers

-0.32

 distrik

-0.31

IBOutlet

-0.29

kreises

-0.29

Gesellschaft

-0.28

 Савезне

-0.28

摇摇头

-0.28

POSITIVE LOGITS

 Tenn

0.73

Tenn

0.71

 nakalista

0.67

 Tend

0.66

addContainerGap

0.66

NESSEE

0.65

 Tenor

0.61

 Tensor

0.61

 Tender

0.60

Tend

0.60

Activations Density 0.157%

the term "Ten" and its variations, suggesting a focus on numerical or measurement-related contexts

No Comments

No Known Activations