INDEX

Explanations

questions and inquiries regarding actions or opinions

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 disambiguazione

-0.81

sidemargin

-0.63

 kyse

-0.59

verwijspagina

-0.59

 BorderSide

-0.58

FunctionFlags

-0.56

atschapp

-0.54

 toalha

-0.54

Diwedd

-0.54

ьажоргаш

-0.54

POSITIVE LOGITS

 what

0.92

 WHAT

0.82

what

0.81

 What

0.79

What

0.77

WHAT

0.71

 whats

0.65

何を

0.63

whats

0.58

hva

0.57

Activations Density 0.054%

questions and inquiries regarding actions or opinions

No Comments

No Known Activations