INDEX

Explanations

political terms and references

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L2_Hcat_z_lr1.20e-03_l11.00e+00_ds24576_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v4.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

float32

Hook Name

blocks.2.attn.hook_z

Hook Layer

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.07

1:0.06

2:0.04

3:0.05

4:0.03

5:0.07

6:0.06

7:0.09

8:0.06

9:0.12

10:0.22

11:0.09

Negative Logits

 Chosen

-1.16

 Yamato

-1.00

 beware

-0.94

 Saiyan

-0.91

iage

-0.90

 Chocobo

-0.88

 samurai

-0.88

 pits

-0.87

 Blossom

-0.87

 Located

-0.86

POSITIVE LOGITS

 archive

1.36

 Pastebin

1.31

 Scrib

1.27

 publication

1.27

 quotations

1.25

 archives

1.21

 editor

1.17

Journal

1.17

Quotes

1.14

 Editorial

1.13

Activations Density 0.744%

political terms and references

No Comments

No Known Activations