INDEX

Explanations

references to conflict, particularly in political or military contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.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

RegistryLite

-0.39

 litté

-0.38

География

-0.36

̀

-0.35

 безо

-0.34

 dolce

-0.34

̈́

-0.34

schul

-0.33

 Flags

-0.33

彗

-0.32

POSITIVE LOGITS

 attempt

0.73

 attempts

0.68

 attempting

0.66

 attempted

0.59

attempt

0.59

 tentativo

0.58

 попы

0.57

ApiProperty

0.54

Attempt

0.53

 Attempt

0.53

Activations Density 0.492%

references to conflict, particularly in political or military contexts

No Comments

No Known Activations