INDEX

Explanations

geographical references to islands and locations

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

2:0.06

3:0.04

4:0.06

5:0.03

6:0.08

7:0.39

8:0.03

9:0.04

10:0.11

11:0.07

Negative Logits

oeuv

-1.78

 effic

-1.61

icipated

-1.55

naissance

-1.55

ressive

-1.53

 pestic

-1.49

 curric

-1.48

 workload

-1.47

ertation

-1.47

 Everyday

-1.46

POSITIVE LOGITS

 Guam

1.72

 Guantanamo

1.62

Lim

1.59

areth

1.58

 Switzerland

1.52

 Madagascar

1.51

 Yugoslavia

1.51

 Antarctica

1.51

 Papua

1.49

near

1.48

Activations Density 0.007%

geographical references to islands and locations

No Comments

No Known Activations