INDEX

Explanations

attends to code-related tokens from textual or coding directive tokens

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

0:0.01

1:0.02

2:0.05

3:0.66

4:0.03

5:0.06

6:0.07

7:0.05

Negative Logits

DockStyle

-0.59

 gddr

-0.53

+#+#

-0.48

qrt

-0.45

principalTable

-0.44

 Pamph

-0.43

 Ecclesiastical

-0.43

verwijspagina

-0.43

MigrationBuilder

-0.43

 Whigs

-0.42

POSITIVE LOGITS

the

0.37

0.32

 some

0.28

its

0.28

all

0.27

of

0.27

and

0.27

may

0.25

 most

0.25

Activations Density 0.015%

attends to code-related tokens from textual or coding directive tokens

No Comments

No Known Activations