INDEX

Explanations

phrases indicative of competitive sports and team dynamics

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

0:0.02

1:0.01

2:0.03

3:0.03

4:0.05

5:0.26

6:0.01

7:0.02

8:0.02

9:0.46

10:0.01

11:0.02

Negative Logits

unci

-2.38

senal

-2.31

apest

-2.20

ワン

-2.01

umbn

-1.98

agos

-1.93

uder

-1.87

ファ

-1.87

 Begin

-1.84

KNOWN

-1.84

POSITIVE LOGITS

 helps

5.49

 proves

5.39

 undermines

5.30

 ensures

5.24

 constitutes

5.21

isn

5.20

 begs

5.08

 contributes

5.07

 improves

5.05

 doesn

5.04

Activations Density 1.196%

phrases indicative of competitive sports and team dynamics

No Known Activations