INDEX

Explanations

punctuation and conjunctions that indicate relationships between ideas or lists

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 9-gemmascope-res-16k

Configuration

google/gemma-scope-9b-it-res/layer_9/width_16k/average_l0_88

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.9.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

Personensuche

-0.52

 defaultstate

-0.48

PROV

-0.44

 Reyes

-0.43

amp

-0.41

只

-0.41

 nime

-0.41

 Kohn

-0.41

 name

-0.40

 Mangel

-0.40

POSITIVE LOGITS

 ſever

0.63

 ſta

0.63

 myſelf

0.60

 tranſ

0.59

ſelf

0.57

 itſelf

0.56

 faſt

0.56

 themſelves

0.53

 pleaſure

0.53

 ſeveral

0.52

Activations Density 0.334%

punctuation and conjunctions that indicate relationships between ideas or lists

No Comments

No Known Activations