INDEX

Explanations

identifiers and references related to user subscriptions

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

thing identifying id, user turn start

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

AndEndTag

-0.65

 menny

-0.39

 Administrativna

-0.39

WithIOException

-0.38

ForValue

-0.38

__*/

-0.37

 FormBuilder

-0.37

intahan

-0.36

XmlAccessorType

-0.36

RuleContext

-0.36

POSITIVE LOGITS

 hepatitis

0.44

ToolStripButton

0.43

 hydrogen

0.42

 disponibilités

0.41

şört

0.40

complexContent

0.40

raux

0.40

 Hydrogen

0.40

 cosmetics

0.40

 Brenn

0.40

Activations Density 0.018%

identifiers and references related to user subscriptions

thing identifying id, user turn start

No Comments

No Known Activations