INDEX

Explanations

numerical or symbolic expressions

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

0:0.21

1:0.19

2:0.04

3:0.05

4:0.04

5:0.12

6:0.04

7:0.02

8:0.07

9:0.07

10:0.05

11:0.04

Negative Logits

 compr

-2.13

 spons

-1.96

 scheduling

-1.86

 regul

-1.84

 booking

-1.83

 licensing

-1.82

 licens

-1.79

 flair

-1.78

 ransom

-1.75

 disciplinary

-1.74

POSITIVE LOGITS

Ann

2.27

Poor

2.18

Mut

2.17

Sy

2.15

Public

2.15

Doc

2.13

Pat

2.12

Syn

2.11

HM

2.11

Or

2.10

Activations Density 0.011%

numerical or symbolic expressions

No Known Activations