INDEX

Explanations

references to significant organizations, events, and individuals in various contexts

New Auto-Interp

Top Features by Cosine Similarity

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L10_Hcat_z_lr1.20e-03_l11.30e+00_ds24576_bs4096_dc1.00e-05_rsanthropic_rie25000_nr4_v9.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

float32

Hook Name

blocks.10.attn.hook_z

Hook Layer

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Activation Function

relu

Embeds

IFrame

Link

Head Attr Weights

0:0.03

1:0.03

2:0.02

3:0.04

4:0.05

5:0.05

6:0.02

7:0.04

8:0.02

9:0.15

10:0.02

11:0.47

Negative Logits

��

-2.40

 purposes

-2.29

rative

-1.96

 Parties

-1.96

 responders

-1.94

veyard

-1.91

 electors

-1.88

 occasions

-1.88

abouts

-1.84

 halftime

-1.82

POSITIVE LOGITS

 newest

2.47

new

2.40

 technologies

2.28

 innovations

2.25

 proposal

2.25

 redesigned

2.23

 technology

2.20

new

2.18

 format

2.17

 bombshell

2.15

Activations Density 0.285%

references to significant organizations, events, and individuals in various contexts

No Known Activations