INDEX

Explanations

phrases indicating quantity and recommendations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.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

 kaikk

-0.40

 only

-0.38

 seluruh

-0.35

 tüm

-0.34

all

-0.34

 Toute

-0.33

 تمامی

-0.32

 entire

-0.32

 kaikki

-0.31

 모든

-0.30

POSITIVE LOGITS

 quelques

0.89

posedge

0.84

 einige

0.81

Quelques

0.81

 trochu

0.80

੝

0.80

 Quelques

0.79

 néhány

0.78

 nicer

0.78

 enkele

0.77

Activations Density 0.521%

phrases indicating quantity and recommendations

No Comments

No Known Activations