INDEX

Explanations

statements that emphasize reasons or explanations for particular phenomena

New Auto-Interp

Top Features by Cosine Similarity

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L3_Hcat_z_lr1.20e-03_l19.00e-01_ds24576_bs4096_dc1.00e-06_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.3.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

1:0.05

2:0.04

3:0.03

4:0.03

5:0.34

6:0.05

7:0.04

8:0.06

9:0.07

10:0.06

11:0.15

Negative Logits

 wagon

-1.76

 descended

-1.73

 barn

-1.73

 stret

-1.72

workshop

-1.71

 duly

-1.70

liga

-1.69

 bapt

-1.68

 favourites

-1.65

ocamp

-1.64

POSITIVE LOGITS

 outweigh

2.45

pmwiki

2.27

advertising

2.13

 outwe

2.08

itiveness

2.07

 answ

2.05

Reviewer

2.00

◼

1.94

 misunderstand

1.89

 ambiguity

1.83

Activations Density 0.052%

statements that emphasize reasons or explanations for particular phenomena

No Known Activations