INDEX

Explanations

attends to tokens that relate to language or textual elements from tokens that signify punctuation or other non-word characters

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

1:0.04

2:0.08

3:0.13

4:0.14

5:0.04

6:0.06

7:0.43

Negative Logits

-1.04

-0.95

in

-0.93

-0.86

<eos>

-0.84

to

-0.84

-0.83

and

-0.82

“

-0.81

↵↵

-0.80

POSITIVE LOGITS

 myſelf

1.86

 itſelf

1.84

 Monfieur

1.80

 Anſ

1.74

 purpoſe

1.74

Efq

1.73

 pleaſure

1.69

 Reſ

1.66

 ―――――

1.65

 ་་

1.63

Activations Density 0.017%

attends to tokens that relate to language or textual elements from tokens that signify punctuation or other non-word characters

No Comments

No Known Activations