INDEX

Explanations

attends to birational-related tokens from related terms within the same context

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

0:0.03

1:0.10

2:0.12

3:0.06

4:0.11

5:0.09

6:0.26

7:0.19

Negative Logits

 Monfieur

-2.89

 myſelf

-2.80

 itſelf

-2.61

 pleaſure

-2.59

 Theſe

-2.55

Efq

-2.53

 Jefus

-2.50

 ―――――

-2.47

 Majefty

-2.42

ſelf

-2.36

POSITIVE LOGITS

1.40

1.27

1.19

in

1.12

1.11

1.09

“

1.09

1.07

1.05

Activations Density 0.004%

attends to birational-related tokens from related terms within the same context

No Comments

No Known Activations