INDEX

Explanations

references to currency and mathematical variables

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

€ and epsilon

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

__":

-0.43

under

-0.43

</i>

-0.40

-0.39

agrave

-0.38

</b>

-0.37

tabular

-0.37

</h2>

-0.36

</h3>

-0.35

POSITIVE LOGITS

€

1.15

€

1.02

 (€

1.01

€)

0.88

 €€

0.83

 €/

0.83

 €,

0.80

URLException

0.79

€/

0.79

 euros

0.78

Activations Density 0.009%

references to currency and mathematical variables

€ and epsilon

No Comments

No Known Activations