INDEX

Explanations

attends to verbs expressing action or states from corresponding tokens denoting similar verbs

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

0:0.10

1:0.16

2:0.08

3:0.11

4:0.28

5:0.12

6:0.05

7:0.05

Negative Logits

"})

-0.29

"}>

-0.28

Biographie

-0.27

."],

-0.26

pyx

-0.26

 propOrder

-0.26

isieren

-0.25

 McClure

-0.25

pira

-0.25

POSITIVE LOGITS

BASEPATH

0.38

 Akismet

0.36

GraphicsUnit

0.35

 AssemblyCulture

0.32

 heapq

0.32

 ComVisible

0.32

 AssemblyTitle

0.32

 Rump

0.31

DataAnnotations

0.30

Bref

0.30

Activations Density 0.184%

attends to verbs expressing action or states from corresponding tokens denoting similar verbs

No Comments

No Known Activations