INDEX

Explanations

years represented in the format '1980s', '1980s', '1980's', '1980' and the word '1990s'

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

0:0.04

1:0.02

2:0.19

3:0.06

4:0.27

5:0.06

6:0.02

7:0.02

8:0.04

9:0.14

10:0.06

11:0.02

Negative Logits

YP

-1.44

odox

-1.43

rawdownloadcloneembedreportprint

-1.41

eton

-1.38

lain

-1.32

 srfAttach

-1.31

ardless

-1.28

idon

-1.26

Lear

-1.25

VPN

-1.23

POSITIVE LOGITS

ablishment

1.62

glers

1.59

rpm

1.54

Jah

1.45

 Recession

1.40

 onwards

1.39

ierre

1.36

 revolving

1.34

 nostalg

1.34

bler

1.34

Activations Density 0.004%

years represented in the format '1980s', '1980s', '1980's', '1980' and the word '1990s'

No Comments

No Known Activations