INDEX

Explanations

attends to mathematical notation or codes from tokens indicating related concepts or operations

oai_attention-head · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 0-gemmascope-att-16k

Configuration

google/gemma-scope-2b-pt-att/layer_0/width_16k/average_l0_104

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.0.attn.hook_z

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.62

1:0.02

2:0.01

3:0.01

4:0.19

5:0.02

6:0.05

7:0.04

Negative Logits

`;

-0.40

-0.32

%";

-0.29

":

-0.28

</i>

-0.27

*/

-0.27

例文帳に追加

-0.27

`).

-0.27

"]

-0.26

POSITIVE LOGITS

 myſelf

0.49

 Monfieur

0.46

 pleaſure

0.46

 itſelf

0.44

 Cæsar

0.44

principalTable

0.44

 purpoſe

0.44

 Reſ

0.43

 Anſ

0.43

 Majefty

0.42

Activations Density 0.085%

attends to mathematical notation or codes from tokens indicating related concepts or operations

No Comments

No Known Activations