INDEX

Explanations

attends to tokens representing legal roles or positions from tokens that are references, such as names or titles, in legal contexts

oai_attention-head · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 10-gemmascope-att-16k

Configuration

google/gemma-scope-2b-pt-att/layer_10/width_16k/average_l0_70

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.10.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

0:0.14

1:0.21

2:0.15

3:0.06

4:0.09

5:0.04

6:0.09

7:0.18

Negative Logits

 EconPapers

-0.34

MLLoader

-0.27

CopyWith

-0.27

最快更新

-0.27

 CURIAM

-0.27

SizeF

-0.26

AndEndTag

-0.26

arwal

-0.25

ortive

-0.25

PDOException

-0.25

POSITIVE LOGITS

InSection

0.26

 nahilalakip

0.25

LookAnd

0.24

ApiOperation

0.24

Ци

0.24

yntaxException

0.24

ис

0.24

kan

0.24

 JAXBElement

0.24

 متعلقه

0.23

Activations Density 0.213%

attends to tokens representing legal roles or positions from tokens that are references, such as names or titles, in legal contexts

No Comments

No Known Activations