INDEX

Explanations

syntax and structure related to programming or code snippets

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

code program 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

 Italijani

-0.64

AsUp

-0.60

TintMode

-0.59

 BoxFit

-0.58

发表于

-0.56

NameInMap

-0.55

 تضيفلها

-0.54

yarnpkg

-0.52

ThroughAttribute

-0.51

writeFieldEnd

-0.51

POSITIVE LOGITS

Personendaten

0.42

 wartawan

0.40

ResumeLayout

0.39

 chegada

0.38

 bekende

0.38

 yaşında

0.38

 modalità

0.38

omal

0.38

 olup

0.38

 wład

0.37

Activations Density 0.022%

syntax and structure related to programming or code snippets

code program identifiers

No Comments

No Known Activations