INDEX

Explanations

references to emotions or reactions within social contexts

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_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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.02

1:0.02

2:0.08

3:0.06

4:0.21

5:0.03

6:0.18

7:0.12

8:0.03

9:0.07

10:0.06

11:0.07

Negative Logits

 flashlight

-1.24

 ALWAYS

-1.19

 masturb

-1.19

Shift

-1.14

 foreskin

-1.13

Cola

-1.12

istries

-1.12

 guided

-1.11

 recip

-1.11

 successfully

-1.10

POSITIVE LOGITS

 Horowitz

1.39

 Colomb

1.36

 Santos

1.36

 Byrne

1.26

 Brock

1.25

 Vale

1.23

igans

1.21

 Regina

1.21

ablishment

1.19

Kop

1.19

Activations Density 0.001%

references to emotions or reactions within social contexts

No Comments

No Known Activations