INDEX

Explanations

numerical values related to collective matters

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.06

1:0.04

2:0.18

3:0.08

4:0.16

5:0.07

6:0.02

7:0.04

8:0.06

9:0.15

10:0.05

11:0.03

Negative Logits

 trail

-1.13

nox

-1.13

DPR

-1.12

 Grill

-1.09

amba

-1.09

 mercury

-1.08

 cruise

-1.06

 mist

-1.03

pson

-1.01

 Slim

-1.01

POSITIVE LOGITS

Unity

1.43

 Cosponsors

1.43

atures

1.38

iations

1.35

isations

1.30

isms

1.26

Publisher

1.23

ATIONS

1.22

together

1.21

hesive

1.21

Activations Density 0.010%

numerical values related to collective matters

No Comments

No Known Activations