INDEX

Explanations

attends to tokens labeled "neg" or "pos" from tokens labeled "Su" or "the."

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.10

1:0.11

2:0.12

3:0.23

4:0.09

5:0.05

6:0.13

7:0.12

Negative Logits

درسة

-0.28

ergies

-0.28

WebMethod

-0.27

katapos

-0.27

 című

-0.26

CopyWith

-0.26

ssysteme

-0.26

municipi

-0.25

لاثة

-0.25

Combien

-0.25

POSITIVE LOGITS

IsMutable

0.29

 Infórmanos

0.29

Билгалдахарш

0.28

">//

0.28

="#"><

0.28

 <<<<<<<<<<<<<<

0.27

 nawr

0.27

 ligiloj

0.27

muer

0.27

endphp

0.27

Activations Density 0.002%

attends to tokens labeled "neg" or "pos" from tokens labeled "Su" or "the."

No Comments

No Known Activations