INDEX

Explanations

attends to tokens indicating conditions or contexts from subsequent tokens that specify those conditions

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

0:0.34

1:0.31

2:0.08

3:0.06

4:0.04

5:0.02

6:0.03

7:0.08

Negative Logits

enterOuterAlt

-0.28

nergy

-0.28

astrar

-0.27

nomin

-0.26

Cas

-0.26

Car

-0.26

змом

-0.25

 parte

-0.25

homonymie

-0.25

POSITIVE LOGITS

 betweenstory

0.48

setViewportView

0.41

MigrationBuilder

0.40

 protoimpl

0.39

__':

0.38

 oprot

0.36

 jspb

0.36

LookAnd

0.36

InitVars

0.36

 mergeFrom

0.36

Activations Density 0.481%

attends to tokens indicating conditions or contexts from subsequent tokens that specify those conditions

No Comments

No Known Activations