INDEX

Explanations

references to friends and relationships

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

Follows sentence endings or the start of turns

np_acts-logits-general · gemini-2.0-flash

positive evaluation of friends' work

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.20.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

 Gatwick

-0.38

 propOrder

-0.35

ItemBackground

-0.35

IContainer

-0.34

坞

-0.34

 StatelessWidget

-0.34

WithFormat

-0.33

 EconPapers

-0.33



-0.32

 onPause

-0.32

POSITIVE LOGITS

 wonderful

0.58

 amazing

0.56

excellent

0.56

incredible

0.56

 wonderfully

0.56

amazing

0.56

 incredible

0.55

wonderful

0.55

 excellent

0.53

 fantastic

0.53

Activations Density 0.244%

references to friends and relationships

Follows sentence endings or the start of turns

positive evaluation of friends' work

No Comments

No Known Activations