INDEX

Explanations

attends to "results" from related tokens concerning studies and findings

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-att/layer_0/width_16k/average_l0_61

Prompts (Dashboard)

16,384 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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.01

1:0.02

2:0.60

3:0.01

4:0.00

5:0.00

6:0.04

7:0.04

8:0.00

9:0.00

10:0.00

11:0.00

12:0.01

13:0.09

14:0.08

15:0.01

Negative Logits

 enfans

-0.50

 varandra

-0.46

 avoient

-0.42

 miniaturka

-0.41

 stället

-0.40

 auroit

-0.40

 feroit

-0.40

 desierto

-0.39

 Italij

-0.39

 démocr

-0.38

POSITIVE LOGITS

ren

0.42

uar

0.42

zac

0.42

rol

0.42

Moj

0.41

uj

0.41

taj

0.41

inum

0.41

olt

0.41

atet

0.40

Activations Density 0.092%

attends to "results" from related tokens concerning studies and findings

No Comments

No Known Activations