INDEX

Explanations

descriptive and emotionally charged adjectives

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 10-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_10/width_16k/average_l0_113

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.10.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Personensuche

-0.46

bewerken

-0.45

 JpaRepository

-0.43

 Allociné

-0.41

 pinulongan

-0.41

FromArgb

-0.41

 wartawan

-0.40

uttgart

-0.39

glGen

-0.38

 Empieza

-0.38

POSITIVE LOGITS

 lovely

0.63

 charming

0.61

tagHelperRunner

0.58

 ErrIntOverflow

0.56

 elegant

0.56

lovely

0.54

 impressive

0.53

Lovely

0.52

 informative

0.52

 frumo

0.51

Activations Density 0.860%

descriptive and emotionally charged adjectives

No Comments

No Known Activations