INDEX

Explanations

discussions of religious contradictions and the nature of faith

oai_token-act-pair · gpt-4o-mini Triggered by @bot

attends to the token "Jesus" from subsequent pronouns or words related to Jesus.

oai_attention-head · claude-3-5-sonnet-20240620 Triggered by @danielepoterti

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-att-kk

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L9_Hcat_z_lr1.20e-03_l11.20e+00_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.9.attn.hook_z

Hook Layer

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Activation Function

relu

Embeds

IFrame

Link

•Religon

No Comments

Head Attr Weights

0:0.13

1:0.02

2:0.11

3:0.10

4:0.03

5:0.11

6:0.03

7:0.08

8:0.26

9:0.03

10:0.03

11:0.02

Negative Logits

zai

-3.63

 Islamabad

-3.56

 Karachi

-3.55

-3.13

 Algeria

-3.07

 Uzbek

-3.06

 Finnish

-2.89

 サーティワン

-2.88

 Pakistani

-2.85

Pakistan

-2.75

POSITIVE LOGITS

 Jesus

5.57

Jesus

5.41

 Galile

5.09

 crucifix

4.99

 Messiah

4.73

 Gospel

4.68

 disciples

4.66

 Bethlehem

4.50

 apostles

4.40

 Christ

4.38

Activations Density 0.068%

discussions of religious contradictions and the nature of faith

attends to the token "Jesus" from subsequent pronouns or words related to Jesus.

No Comments

No Known Activations