INDEX

Explanations

attends to tokens with specific class and value attributes from tokens indicating a lack of surrounding context with class or parameter attributes

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

0:0.05

1:0.03

2:0.02

3:0.03

4:0.66

5:0.05

6:0.07

7:0.04

Negative Logits

 Anſ

-0.40

 Reſ

-0.37

 purpoſe

-0.34

 Diſ

-0.34

 itſelf

-0.34

 Houſe

-0.33

^(@)

-0.33

 Theſe

-0.32

 Inscrivez

-0.32

 ſeveral

-0.32

POSITIVE LOGITS

 tartalomajánló

0.45

tagHelperRunner

0.45

mybatisplus

0.39

@"/

0.39

 NSCoder

0.39

migrationBuilder

0.39

localctx

0.38

 насељу

0.37

:✨

0.36

AccessorTable

0.36

Activations Density 0.024%

attends to tokens with specific class and value attributes from tokens indicating a lack of surrounding context with class or parameter attributes

No Comments

No Known Activations