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

1:0.05

2:0.07

3:0.07

4:0.07

5:0.09

6:0.07

7:0.08

8:0.08

9:0.09

10:0.10

11:0.08

Negative Logits

espie

-1.71

izo

-1.64

raints

-1.61

Unit

-1.60

peg

-1.56

 dosage

-1.51

Kle

-1.50

ibu

-1.50

etooth

-1.49

went

-1.46

POSITIVE LOGITS

 Truth

1.93

 Consortium

1.80

 Correspond

1.58

Truth

1.48

 Junk

1.46

 bask

1.45

 Salon

1.44

justice

1.44

 Scarlet

1.42

ocrats

1.41

Activations Density 0.000%

No Known Activations

This feature has no known activations.