INDEX

Explanations

references to processes and actions described as "things."

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

1:0.03

2:0.12

3:0.09

4:0.20

5:0.09

6:0.08

7:0.12

8:0.06

9:0.03

10:0.06

11:0.05

Negative Logits

enter

-1.47

 contains

-1.40

 partName

-1.37

>=

-1.37

 enters

-1.35

 characteristic

-1.34

 involuntary

-1.29

 feature

-1.29

 Contains

-1.27

*/(

-1.26

POSITIVE LOGITS

 theirs

1.58

ilaterally

1.48

Policy

1.39

reated

1.35

aunted

1.33

SPONSORED

1.32

ierrez

1.31

 McCain

1.29

 Podesta

1.28

 hers

1.26

Activations Density 0.002%

references to processes and actions described as "things."

No Comments

No Known Activations