INDEX

Explanations

programming terminology and code structure

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

user code starts

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

évaluateur

-0.58

HomeAsUpEnabled

-0.55

Примітки

-0.45

 nakalista

-0.45

 plufieurs

-0.44

AnchorStyles

-0.44

styleType

-0.44

ésult

-0.44

 ainfi

-0.44

-------------</

-0.44

POSITIVE LOGITS

ocino

0.37

'{@

0.36

 variables

0.35

eneuve

0.33

RestController

0.32

ciado

0.32

 implements

0.32

 vari

0.32

 AssemblyTitle

0.32

isUser

0.31

Activations Density 0.005%

programming terminology and code structure

user code starts

No Comments

No Known Activations