INDEX

Explanations

attends to the token "frequent" from the tokens "use" and "used."

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

0:0.30

1:0.02

2:0.02

3:0.02

4:0.42

5:0.09

6:0.02

7:0.07

Negative Logits

 Roskov

-0.92

^(@)

-0.85

 ddelweddau

-0.84

Geplaatst

-0.83

HtmlAttribute

-0.82

FTFY

-0.79

esterday

-0.79

 للاسماء

-0.78

 صوتيه

-0.78

 tartalomajánló

-0.78

POSITIVE LOGITS

<eos>

0.47

↵↵

0.45

0.44

0.43

0.42

0.41

Activations Density 0.061%

attends to the token "frequent" from the tokens "use" and "used."

No Comments

No Known Activations