INDEX

Explanations

phrases related to conflicts or disputes

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L2_Hcat_z_lr1.20e-03_l11.00e+00_ds24576_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v4.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

float32

Hook Name

blocks.2.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

0:0.03

1:0.11

2:0.05

3:0.10

4:0.07

5:0.10

6:0.09

7:0.04

8:0.08

9:0.15

10:0.08

11:0.03

Negative Logits

 Hawai

-1.48

 ______

-1.44

edience

-1.43

colored

-1.31

 someday

-1.31

 Harlem

-1.31

 Chattanooga

-1.30

 coloring

-1.30

 distilled

-1.29

 gotten

-1.28

POSITIVE LOGITS

 apologise

1.65

hend

1.64

 tonnes

1.62

 centre

1.59

 emphas

1.58

 realise

1.57

 apologised

1.56

 Surrey

1.55

 Trident

1.55

 aluminium

1.53

Activations Density 0.075%

phrases related to conflicts or disputes

No Comments

No Known Activations