INDEX

Explanations

numeric values or identifiers, particularly in a structured data context

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

hexadecimal identifiers

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

存于互联网档案馆

-0.90

ſelf

-0.84

 surla

-0.83

ſelves

-0.80

Efq

-0.79

<unused23>

-0.77

[@BOS@]

-0.77

<unused14>

-0.77

<unused8>

-0.77

<unused16>

-0.77

POSITIVE LOGITS

好

0.30

0.29

ं

0.28

À

0.27

越

0.27

‍

0.26

母

0.26

より

0.25

</u>

0.25

0.24

Activations Density 0.006%

numeric values or identifiers, particularly in a structured data context

hexadecimal identifiers

No Comments

No Known Activations