INDEX

Explanations

numerical patterns, particularly focusing on IP addresses and return codes

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

 bordado

-0.41

occasione

-0.40

 bordir

-0.39

 Constitucional

-0.38

 zanah

-0.37

 bordados

-0.37

 Khusus

-0.37

 larmes

-0.36

 barnen

-0.35

 hidráulica

-0.35

POSITIVE LOGITS

 للمعارف

0.63

tvguidetime

0.60

styleType

0.57

Portail

0.57

addGap

0.54

Jîn

0.51

LookAnd

0.50

setVerticalGroup

0.50

stdafx

0.50

 clavic

0.49

Activations Density 0.011%

numerical patterns, particularly focusing on IP addresses and return codes

No Comments

No Known Activations