INDEX

Explanations

words or names associated with notable individuals or entities in sports

New Auto-Interp

Top Features by Cosine Similarity

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L5_Hcat_z_lr1.20e-03_l11.00e+00_ds49152_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v9.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

Data Type

float32

Hook Name

blocks.5.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

1:0.03

2:0.05

3:0.06

4:0.04

5:0.05

6:0.44

7:0.04

8:0.05

9:0.07

10:0.07

11:0.04

Negative Logits

urdue

-1.28

Spoiler

-1.17

 dign

-1.17

 shortest

-1.16

 humble

-1.13

 trim

-1.13

 role

-1.11

 territorial

-1.11

 liberty

-1.10

BILITY

-1.10

POSITIVE LOGITS

opal

1.44

ographs

1.44

hov

1.42

obin

1.39

agate

1.38

ohm

1.34

��

1.34

zinski

1.32

chrom

1.32

uden

1.31

Activations Density 0.021%

words or names associated with notable individuals or entities in sports

No Known Activations