INDEX

Explanations

terms related to features and specifications of products or systems

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

equipped with features

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

RegressionTest

-0.63

HostException

-0.59

DockStyle

-0.56

曖昧さ回避

-0.56

adaptiveStyles

-0.54

memoized

-0.54

PickerController

-0.53

complexContent

-0.51

fastjson

-0.49

SpringBootTest

-0.46

POSITIVE LOGITS

 posiada

0.44

 dodatk

0.44

 dodatkowe

0.44

 silicona

0.43

 dilengkapi

0.42

 fitur

0.42

 provide

0.42

 zawiera

0.42

 adjustable

0.41

 wyposaż

0.41

Activations Density 0.036%

terms related to features and specifications of products or systems

equipped with features

No Comments

No Known Activations