INDEX

Explanations

extensions of words related to judgment or evaluation

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

accumulating suffixes

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

vœ

-0.39

 professeur

-0.31

 lendemain

-0.30

 comunicarse

-0.29

:✨

-0.29

 désert

-0.28

 Hivatalos

-0.28

pecabe

-0.28

 Lainnya

-0.27

 acá

-0.27

POSITIVE LOGITS

</>

0.72

 resourceCulture

0.65

the

0.64

 ddelwed

0.62

 メンテナ

0.61

☸

0.61



0.60

𖧵

0.60



0.60

 ſuo

0.60

Activations Density 0.021%

extensions of words related to judgment or evaluation

accumulating suffixes

No Comments

No Known Activations