INDEX

Explanations

No Explanations Found

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

0:0.08

1:0.09

2:0.07

3:0.07

4:0.08

5:0.07

6:0.08

7:0.07

8:0.10

9:0.08

10:0.07

11:0.07

Negative Logits

osate

-1.70

 laun

-1.67

ngth

-1.64

 gobl

-1.58

INF

-1.53

 surpassed

-1.52

raltar

-1.51

 subreddit

-1.47

DoS

-1.46

ibo

-1.46

POSITIVE LOGITS

 respons

1.74

sheets

1.73

 bargain

1.58

 prosecut

1.56

 eleg

1.51

 autos

1.50

erred

1.44

 Fifth

1.44

dress

1.43

 barg

1.43

Activations Density 0.000%

No Known Activations

This feature has no known activations.