INDEX

Explanations

attends to numerical tokens represented in both word form and numerical form from numerical quantities that follow and precede them in the context

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

0:0.10

1:0.03

2:0.03

3:0.04

4:0.07

5:0.02

6:0.07

7:0.60

Negative Logits

-0.51

-0.50

So

-0.48

«

-0.47

‘

-0.47

“

-0.46

-0.45

POSITIVE LOGITS

 purpoſe

0.91

 Reſ

0.91

 greateſt

0.90

 pleaſure

0.89

 Monfieur

0.88

 Diſ

0.85

 Anſ

0.84

 faſt

0.82

 ſmall

0.82

 diſt

0.80

Activations Density 0.084%

attends to numerical tokens represented in both word form and numerical form from numerical quantities that follow and precede them in the context

No Comments

No Known Activations