INDEX

Explanations

phrases that denote possession or describe ownership

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

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

Not in Any Lists

No Comments

Head Attr Weights

0:0.01

1:0.01

2:0.11

3:0.06

4:0.16

5:0.03

6:0.08

7:0.29

8:0.03

9:0.03

10:0.08

11:0.06

Negative Logits

龍契士

-1.95

autions

-1.82

 advoc

-1.72

quished

-1.68

aspberry

-1.65

aturdays

-1.64

iolet

-1.63

uesday

-1.63

cffff

-1.61

conservancy

-1.58

POSITIVE LOGITS

 faulty

1.89

 error

1.56

 brunt

1.45

 fragment

1.44

 infringing

1.36

 Aston

1.34

 variable

1.33

 mistake

1.31

 theor

1.30

 leak

1.29

Activations Density 0.002%

phrases that denote possession or describe ownership

No Comments

No Known Activations