INDEX

Explanations

attends to the token "this" from identical tokens that follow or similar constructs

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

0:0.07

1:0.21

2:0.05

3:0.06

4:0.10

5:0.37

6:0.05

7:0.05

Negative Logits

 Tutu

-0.45

BagConstraints

-0.41

ัติ

-0.40

 AssemblyTitle

-0.39

 Kome

-0.37

rrggbb

-0.36

ussis

-0.35

krok

-0.35

 tauto

-0.35

 Numer

-0.35

POSITIVE LOGITS

dė

0.33

 HasFactory

0.33

にほんブログ村

0.30

 élect

0.29

ordnet

0.29

ième

0.29

แน่น

0.29

 murs

0.28

 fumée

0.28

словия

0.28

Activations Density 0.030%

attends to the token "this" from identical tokens that follow or similar constructs

No Comments

No Known Activations