INDEX

Explanations

phrases indicating progress or evaluation of performance and results

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

0:0.01

1:0.03

2:0.10

3:0.07

4:0.01

5:0.04

6:0.05

7:0.11

8:0.15

9:0.19

10:0.05

11:0.13

Negative Logits

along

-1.34

owned

-1.33

ussed

-1.33

interrupted

-1.32

depending

-1.17

UME

-1.15

arthed

-1.13

eria

-1.12

DragonMagazine

-1.12

gered

-1.11

POSITIVE LOGITS

DERR

1.24

 negativity

1.20

clus

1.19

otin

1.18

 professionalism

1.18

 inconsistencies

1.17

 honesty

1.16

 assurances

1.14

 sincerity

1.14

 inconsistency

1.13

Activations Density 0.072%

phrases indicating progress or evaluation of performance and results

No Comments

No Known Activations