INDEX

Explanations

features related to skills and abilities that enhance user powers or characteristics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 27-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_27/width_16k/average_l0_118

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.27.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

 informée

-0.48

usermodel

-0.42

 heureux

-0.41

 EconPapers

-0.40

 réfugi

-0.39

 précieux

-0.38

 maîtres

-0.38

onomous

-0.37

seitige

-0.37

таратура

-0.37

POSITIVE LOGITS

 floating

0.65

 materialized

0.58

 vanishing

0.56

 teleport

0.56

 flying

0.56

 glowing

0.55

 disappeared

0.54

 disappears

0.54

 vapor

0.54

 suddenly

0.54

Activations Density 0.269%

features related to skills and abilities that enhance user powers or characteristics

No Comments

No Known Activations