INDEX

Explanations

references to educational institutions and alumni status

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

Category: prefixes

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-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

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

 shaping

-0.34

erschein

-0.32

dek

-0.31

 agujas

-0.31

красивые

-0.30

 rolls

-0.30

 majd

-0.29

 toek

-0.28

schöne

-0.28

打

-0.28

POSITIVE LOGITS

 Biôgrafia

0.64

TagMode

0.64

-------------</

0.62

TypedDataSet

0.62

RectangleBorder

0.60

Хьажоргаш

0.59

writeFieldEnd

0.57

 autorytatywna

0.57

 NSCoder

0.57

 ویکی‌پدی

0.56

Activations Density 0.066%

references to educational institutions and alumni status

Category: prefixes

No Comments

No Known Activations