INDEX

Explanations

questions and discussions around specific research topics and categories

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

topics and areas

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

astify

-0.57

новништво

-0.57

flashdata

-0.52

 fiberglass

-0.50

 ChromeDriver

-0.47

 miſ

-0.47

 myſelf

-0.46

 مشين

-0.45

 escova

-0.44

 pleaſure

-0.43

POSITIVE LOGITS

 topics

0.69

 areas

0.65

 themes

0.61

 Areas

0.61

Areas

0.54

areas

0.54

 AREAS

0.53

topics

0.52

 temáticas

0.51

 Topics

0.51

Activations Density 0.065%

questions and discussions around specific research topics and categories

topics and areas

No Comments

No Known Activations