INDEX

Explanations

references to formal announcements or declarations

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

attribution of statements

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

LLocation

-0.67

 мәкал

-0.58

 OnTrigger

-0.51

 pręd

-0.50

uxxxx

-0.49

 auroit

-0.48

تقاوى

-0.47



-0.47

 faſt

-0.46

BufferException

-0.45

POSITIVE LOGITS

 ujednoznacz

0.47

 speaker

0.45

 speakers

0.43

lwjgl

0.40

 quoted

0.39

 kasarigan

0.39

 supremo

0.38

Kjelder

0.38

 exclaimed

0.37

strerror

0.37

Activations Density 0.015%

references to formal announcements or declarations

attribution of statements

No Comments

No Known Activations