INDEX

Explanations

attends to the closing sequence markers preceding various tokens that indicate legal judgments or commentary

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

0:0.11

1:0.07

2:0.09

3:0.10

4:0.18

5:0.16

6:0.11

7:0.13

Negative Logits

-3.64

-3.53

-3.23

↵↵

-3.19

-3.06

↵

-2.73

-2.72

in

-2.64

-2.61

the

-2.59

POSITIVE LOGITS

 myſelf

6.38

 itſelf

6.16

Efq

6.03

 Monfieur

5.69

 Theſe

5.59

 ―――――

5.47

 pleaſure

5.44

 Jefus

5.44

 Majefty

5.34

 Anſ

5.31

Activations Density 0.009%

attends to the closing sequence markers preceding various tokens that indicate legal judgments or commentary

No Comments

No Known Activations