INDEX

Explanations

attends to terms related to sensing or detection from later tokens that pertain to their respective actions or functionalities

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

0:0.11

1:0.14

2:0.13

3:0.13

4:0.11

5:0.05

6:0.12

7:0.17

Negative Logits

WebVitals

-0.28

gridx

-0.27

NameInMap

-0.27

 gás

-0.25

picasso

-0.25

 heroin

-0.24

 Freuden

-0.24

SpringBoot

-0.24

 incest

-0.23

 claw

-0.23

POSITIVE LOGITS

NUMX

0.28

getSimpleName

0.28

eabouts

0.26

はじめに

0.26

/*++

0.26

Capítulo

0.26

ApiModelProperty

0.25

ptonshire

0.25

চ্

0.25

ctrica

0.24

Activations Density 0.356%

attends to terms related to sensing or detection from later tokens that pertain to their respective actions or functionalities

No Comments

No Known Activations