INDEX

Explanations

single characters or letters that represent certain categories or classifications

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

 Roskov

-0.70

featureID

-0.68

تقاوى

-0.66

parsedMessage

-0.61

AddTagHelper

-0.60

 nakalista

-0.59

WriteBarrier

-0.59

AccessorTable

-0.58

 Мексичка

-0.57

 disambiguazione

-0.57

POSITIVE LOGITS

0.45

ListTile

0.44

Flu

0.43

 patin

0.42

 Dossier

0.42

Dis

0.42

Position

0.42

 Glut

0.41

Package

0.41

 vesti

0.41

Activations Density 0.047%

single characters or letters that represent certain categories or classifications

No Comments

No Known Activations