INDEX

Explanations

attends to tokens marked with [[ ]] that are individually positioned after the "+" token throughout the code snippets

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 10-gemmascope-att-16k

Configuration

google/gemma-scope-2b-pt-att/layer_10/width_16k/average_l0_70

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.10.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

0:0.55

1:0.10

2:0.06

3:0.04

4:0.05

5:0.06

6:0.04

7:0.07

Negative Logits

 kasarigan

-0.45

tanleria

-0.32

ver

-0.31

ruiter

-0.31

 estekak

-0.30

 iNdEx

-0.28

 الحره

-0.28

oneph

-0.28

 tide

-0.28

RequestHandler

-0.28

POSITIVE LOGITS

 Majefty

0.51

 myſelf

0.50

 purpoſe

0.47

 ſtate

0.44

 itſelf

0.43

 Houſe

0.43

 ſeveral

0.43

 poffe

0.42

 ſame

0.41

 themſelves

0.40

Activations Density 0.105%

attends to tokens marked with [[ ]] that are individually positioned after the "+" token throughout the code snippets

No Comments

No Known Activations