INDEX

Explanations

attends to language-related tokens from various production and creative roles

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

0:0.01

1:0.05

2:0.05

3:0.16

4:0.04

5:0.05

6:0.44

7:0.14

Negative Logits

Personensuche

-0.60

TagMode

-0.58

gr

-0.57

 تضيفلها

-0.53

Manbalar

-0.51

 nakalista

-0.51

BeginContext

-0.50

Abitanti

-0.49

Général

-0.48

Collegamenti

-0.47

POSITIVE LOGITS

 Theſe

1.13

 Monfieur

1.04

 Majefty

1.00

Efq

1.00

 itſelf

0.98

 Jefus

0.97

ſelf

0.94

 myſelf

0.93

 Houſe

0.91

 pleaſure

0.90

Activations Density 0.477%

attends to language-related tokens from various production and creative roles

No Comments

No Known Activations