INDEX

Explanations

the word "one" across various contexts

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

Head Attr Weights

0:0.05

1:0.02

2:0.10

3:0.08

4:0.07

5:0.05

6:0.16

7:0.03

8:0.04

9:0.05

10:0.06

11:0.25

Negative Logits

Lur

-2.15

 belonging

-1.99

 occupant

-1.97

 Marriott

-1.95

 contacts

-1.94

 lodged

-1.92

 accommodation

-1.90

 communicating

-1.84

 lodging

-1.83

 Sharing

-1.83

POSITIVE LOGITS

神

1.96

onential

1.89

Percent

1.81

heses

1.80

ivari

1.77

omial

1.76

ァ

1.75

theless

1.74

 lows

1.72

gebra

1.70

Activations Density 0.000%

the word "one" across various contexts

No Known Activations