INDEX

Explanations

attends to tokens associated with "No" from tokens related to "ge," "y," and other suffixes

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

0:0.68

1:0.01

2:0.03

3:0.01

4:0.12

5:0.04

6:0.02

7:0.05

Negative Logits

ora

-0.34

 iload

-0.33

â

-0.32

 leçons

-0.30

<sup>

-0.30

and

-0.30

 espejos

-0.30

as

-0.30

 journaux

-0.29

 kabát

-0.29

POSITIVE LOGITS

 Мексичка

0.51

ratulations

0.43

SEGUIR

0.43

]=>

0.41

DEPORTES

0.40

elung

0.39

 AssemblyTitle

0.39

BASEPATH

0.38

TagMode

0.38

etcode

0.37

Activations Density 0.035%

attends to tokens associated with "No" from tokens related to "ge," "y," and other suffixes

No Comments

No Known Activations