INDEX

Explanations

references to the concept of upgrading in various contexts

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

upgrade to or from

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

labelText

-0.43

testify

-0.42

⋅

-0.42

 ComVisible

-0.41

(()

-0.41

arynx

-0.41

prnewswire

-0.41

Charsets

-0.41

 DataTypes

-0.40

 voelt

-0.40

POSITIVE LOGITS

 upgrade

1.24

 upgrading

1.15

 Upgrade

1.15

upgrade

1.09

 upgrades

1.09

Upgrades

1.07

 upgraded

1.06

 Upgrades

1.03

Upgrade

1.00

 upgra

0.95

Activations Density 0.004%

references to the concept of upgrading in various contexts

upgrade to or from

No Comments

No Known Activations