INDEX

Explanations

attends from tokens that include a hash symbol to tokens that include an empty space or no specific token

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

0:0.76

1:0.01

2:0.01

3:0.01

4:0.11

5:0.01

6:0.03

7:0.02

Negative Logits

 ſeveral

-0.70

 Majefty

-0.69

 myſelf

-0.67

 pleaſure

-0.66

 purpoſe

-0.66

Efq

-0.65

 reaſon

-0.64

 greateſt

-0.64

 ſtate

-0.63

 Monfieur

-0.63

POSITIVE LOGITS

No

0.39

in

0.38

0.36

The

0.35

0.34

on

0.34

In

0.33

So

0.33

Activations Density 0.053%

attends from tokens that include a hash symbol to tokens that include an empty space or no specific token

No Comments

No Known Activations