INDEX

Explanations

attends to a pattern of tokens related to "rd" from tokens related to "ma"

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

0:0.64

1:0.01

2:0.02

3:0.01

4:0.13

5:0.10

6:0.02

7:0.04

Negative Logits

 GenerationType

-0.40

ethene

-0.40

↩

-0.35

MockBean

-0.35

Controllo

-0.35

centralwidget

-0.34

XmlAccessType

-0.34

 թվական

-0.34

pushFollow

-0.33

endforeach

-0.33

POSITIVE LOGITS

DoubleQuotes

0.48

ContentAsync

0.39

Enumerated

0.38

 showers

0.35

providedIn

0.35

gerton

0.35

 iprot

0.33

lıyor

0.32

__[

0.32

GeoNames

0.31

Activations Density 0.005%

attends to a pattern of tokens related to "rd" from tokens related to "ma"

No Comments

No Known Activations