INDEX

Explanations

phrases indicative of conditions or states that may involve evaluation or assessment

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

after certain words

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

Negative Logits

期刊论文

-0.40

arXiv

-0.38

 simplifié

-0.38

Siehe

-0.33

-0.31

SetOutput

-0.31

rungsseite

-0.31

 Industria

-0.31

POSITIVE LOGITS

 Wikimedijinoj

0.72

 rhestr

0.68

RegistryLite

0.67

 AssemblyCulture

0.63

AsUp

0.63

 فريبيس

0.62

FieldBuilder

0.59

 typelib

0.57

 Мексичка

0.57

IndentedString

0.56

Activations Density 0.289%

phrases indicative of conditions or states that may involve evaluation or assessment

after certain words

No Known Activations