INDEX

Explanations

words indicating quantitative or comparative aspects

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

yce

-0.40

botines

-0.40

 spéciale

-0.39

ногда

-0.37

 Gual

-0.36

 Verwendung

-0.36

gorro

-0.35

 siebie

-0.35

 militaires

-0.35

 almohada

-0.34

POSITIVE LOGITS

 للاسماء

0.71

AddTagHelper

0.68



0.67

 OMITBAD

0.66

enterOuterAlt

0.63

 <<<<<<<<<<<<<<

0.59

uxxxx

0.59

setVerticalGroup

0.59

GOTREF

0.59

 CanadaChoose

0.58

Activations Density 0.020%

words indicating quantitative or comparative aspects

No Comments

No Known Activations