INDEX

Explanations

attends to pronouns indicating entities or relationships from corresponding tokens indicating a category or description

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.20

1:0.35

2:0.13

3:0.05

4:0.04

5:0.04

6:0.04

7:0.11

Negative Logits

onAttach

-0.32

✭✭

-0.32

ArgsConstructor

-0.32

RTLU

-0.30

dónde

-0.30

 الحره

-0.30

Étymologie

-0.30

󠁣

-0.30

 Wikispecies

-0.29

ScopeManager

-0.29

POSITIVE LOGITS

Leaky

0.25

atera

0.23

URIComponent

0.23

 "../../../

0.22

 spur

0.22

setopt

0.21

tagext

0.21

 Jealous

0.21

DeleteMapping

0.21

ampa

0.21

Activations Density 0.466%

attends to pronouns indicating entities or relationships from corresponding tokens indicating a category or description

No Comments

No Known Activations