INDEX

Explanations

attends to "On" tokens from corresponding "[[On]]" tokens

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.02

1:0.73

2:0.03

3:0.02

4:0.05

5:0.02

6:0.04

7:0.05

Negative Logits

On

-0.34

On

-0.31

 From

-0.29

 After

-0.29

 With

-0.29

Sur

-0.27

After

-0.27

picasso

-0.27

As

-0.27

By

-0.27

POSITIVE LOGITS

NUMX

0.47

 الحره

0.41

 pleaſure

0.39

 ſhe

0.39

 consultato

0.39

 ſta

0.38

 Inscrivez

0.38

 houſe

0.37

 ſtand

0.37

 myſelf

0.37

Activations Density 0.074%

attends to "On" tokens from corresponding "[[On]]" tokens

No Comments

No Known Activations