INDEX

Explanations

references to puzzles and games

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 36-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_36/width_16k/average_l0_120

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.36.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

<?

-0.40

 Rptr

-0.40

carter

-0.38

MethodManager

-0.36

ToAction

-0.35

stahl

-0.34

 TAGS

-0.34

 sánh

-0.34

getWriter

-0.33

tisseur

-0.33

POSITIVE LOGITS

 puzzle

2.33

 puzzles

2.11

puzzle

1.90

 Puzzle

1.84

Puzzle

1.74

 rompecabezas

1.67

 puzz

1.66

 Puzzles

1.63

 riddle

1.48

 crossword

1.46

Activations Density 0.300%

references to puzzles and games

No Comments

No Known Activations