INDEX

Explanations

references to credible sources and organizations

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_L10_Hcat_z_lr1.20e-03_l11.30e+00_ds24576_bs4096_dc1.00e-05_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.10.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

0:0.03

1:0.02

2:0.03

3:0.04

4:0.07

5:0.50

6:0.02

7:0.03

8:0.02

9:0.14

10:0.02

11:0.04

Negative Logits

 outweigh

-3.17

 lately

-3.10

obin

-2.91

 nowadays

-2.84

 justifies

-2.83

 poses

-2.82

 anytime

-2.77

 assures

-2.64

 demonstrates

-2.64

 unequivocally

-2.51

POSITIVE LOGITS

but

4.26

but

3.74

then

3.31

until

2.97

only

2.94

But

2.90

 only

2.86

 until

2.68

listed

2.63

BUT

2.62

Activations Density 0.286%

references to credible sources and organizations

No Comments

No Known Activations