INDEX

Explanations

constructions that indicate potential actions or features related to products and services

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

references to inventions and their functionalities

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

first-person perspective

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_65k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

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

onAttach

-0.60

 CreateTagHelper

-0.57

دانشنامهٔ

-0.56

èses

-0.55

mord

-0.50

unknownFields

-0.49

ThroughAttribute

-0.48

prüche

-0.47

 tendency

-0.46

 hoàng

-0.46

POSITIVE LOGITS

 even

0.60

 дори

0.58

even

0.57

новниш

0.56

 bahkan

0.54

 successively

0.52

 jopa

0.52

 سكانية

0.51

 Bäume

0.50

 especially

0.50

Activations Density 0.138%

constructions that indicate potential actions or features related to products and services

references to inventions and their functionalities

first-person perspective

No Comments

No Known Activations