INDEX

Explanations

words indicating possibilities and identities

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

New Auto-Interp

Top Features by Cosine Similarity

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L3_Hcat_z_lr1.20e-03_l19.00e-01_ds24576_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v9.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

float32

Hook Name

blocks.3.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

0:0.02

1:0.02

2:0.07

3:0.25

4:0.01

5:0.02

6:0.13

7:0.08

8:0.04

9:0.03

10:0.05

11:0.23

Negative Logits

culosis

-1.69

ahime

-1.29

ioxide

-1.27

��

-1.26

utils

-1.23

metry

-1.22

chwitz

-1.21

 unfocusedRange

-1.19

encing

-1.17

leased

-1.15

POSITIVE LOGITS

 conceivable

1.51

 imaginable

1.46

ieth

1.32

 fraction

1.24

 feasible

1.21

(<

1.21

iest

1.19

 EVER

1.19

erest

1.18

 shin

1.15

Activations Density 0.010%

words indicating possibilities and identities

No Comments

No Known Activations