INDEX

Explanations

repeated conjunctions and quantifiers in mathematical and technical descriptions

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

-0.45

to

-0.37

 instead

-0.36

 only

-0.33

Ju

-0.32

 beyond

-0.31

for

-0.31

if

-0.31

 while

-0.31

 finally

-0.30

POSITIVE LOGITS

 autorytatywna

0.74

 nahilalakip

0.71

aarrggbb

0.70

səhifə

0.68

Personendaten

0.67

pingente

0.66

جوايز

0.66

bootstrapcdn

0.66

localctx

0.65

RotationOrder

0.65

Activations Density 0.720%

repeated conjunctions and quantifiers in mathematical and technical descriptions

No Comments

No Known Activations