INDEX

Explanations

attends to tokens that denote operations or functional commands from tokens that imply an addition or accumulation

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

0:0.07

1:0.14

2:0.08

3:0.04

4:0.13

5:0.40

6:0.04

7:0.06

Negative Logits

ésult

-0.23

⁩

-0.23

 Epist

-0.23

lofen

-0.23

ukai

-0.23

fei

-0.23

 climax

-0.22

 Firms

-0.22

KOT

-0.22

 commerciales

-0.22

POSITIVE LOGITS

expandindo

0.33

gameserver

0.31

Дереккөздер

0.31

 Normdatei

0.31

 createContext

0.28

 חיצוניים

0.28

 "..\..\..\

0.26

 BoxFit

0.26

 unknownFields

0.26

JoinColumn

0.26

Activations Density 0.122%

attends to tokens that denote operations or functional commands from tokens that imply an addition or accumulation

No Comments

No Known Activations