INDEX

Explanations

attends to the term "target" from associated tokens or phrases that describe what is being aimed at

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

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.08

1:0.08

2:0.06

3:0.02

4:0.41

5:0.22

6:0.03

7:0.07

Negative Logits

↵↵

-0.36

<eos>

-0.34

-0.30

$$

-0.29

μ

-0.29

ho

-0.28

po

-0.28

POSITIVE LOGITS

 Jefus

0.68

Efq

0.66

^(@)

0.60

CodedInputStream

0.59

 BoxFit

0.58

PerformLayout

0.58

 doubtnut

0.58

 nahilalakip

0.57

 Majefty

0.57

 becauſe

0.56

Activations Density 0.081%

attends to the term "target" from associated tokens or phrases that describe what is being aimed at

No Comments

No Known Activations