INDEX

Explanations

attends to various expressions of "after" from preceding token contexts

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

1:0.66

2:0.05

3:0.03

4:0.07

5:0.03

6:0.03

7:0.05

Negative Logits

 Theſe

-0.62

 Diſ

-0.61

 leaſt

-0.61

 ſeveral

-0.59

 Chriftian

-0.59

 greateſt

-0.58

 Majefty

-0.58

 Monfieur

-0.58

 auroit

-0.58

 pleaſure

-0.57

POSITIVE LOGITS

0.44

0.40

cu

0.38

“

0.36

un

0.35

the

0.35

app

0.35

‘

0.35

0.34

he

0.34

Activations Density 0.157%

attends to various expressions of "after" from preceding token contexts

No Comments

No Known Activations