INDEX

Explanations

attends to tokens indicating a degree of extent or comparison from subsequent contrasting tokens

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

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.46

1:0.18

2:0.09

3:0.09

4:0.04

5:0.02

6:0.03

7:0.06

Negative Logits

Tikang

-0.45

GEBURTSDATUM

-0.42

 متعلقه

-0.34

 ویکی‌پدیا

-0.33

 propOrder

-0.32

Geplaatst

-0.30



-0.30

 Signalez

-0.30



-0.28

 &___

-0.28

POSITIVE LOGITS

 recourse

0.28

ngdoc

0.28

 Profesional

0.27

 предпо

0.25

 Buxton

0.25

 město

0.24

iją

0.24

 universitario

0.23

 universitaria

0.23

ią

0.23

Activations Density 1.178%

attends to tokens indicating a degree of extent or comparison from subsequent contrasting tokens

No Comments

No Known Activations