INDEX

Explanations

phrases related to someone asking for help or feeling targeted

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

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.02

1:0.02

2:0.29

3:0.17

4:0.11

5:0.02

6:0.04

7:0.05

8:0.05

9:0.04

10:0.09

11:0.07

Negative Logits

 Lighting

-1.70

 Organizations

-1.53

thumbnails

-1.51

 Lans

-1.49

EMS

-1.46

External

-1.45

 Ambro

-1.43

 impact

-1.43

eru

-1.36

agu

-1.35

POSITIVE LOGITS

 lied

1.71

 whore

1.61

!?

1.59

:(

1.58

...?

1.54

huh

1.54

 damned

1.50

#$

1.49

'?

1.48

eeee

1.48

Activations Density 0.142%

phrases related to someone asking for help or feeling targeted

No Comments

No Known Activations