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

Head Attr Weights

0:0.07

1:0.07

2:0.10

3:0.08

4:0.09

5:0.08

6:0.09

7:0.09

8:0.07

9:0.07

10:0.07

11:0.07

Negative Logits

 Sack

-1.75

 Turing

-1.69

 symbolism

-1.66

 continu

-1.61

 Antiqu

-1.58

 Reign

-1.56

 Kard

-1.56

 Thieves

-1.56

 Strongh

-1.55

 stocking

-1.54

POSITIVE LOGITS

odo

1.90

ESA

1.89

itled

1.86

eneg

1.80

ヴァ

1.76

ague

1.73

oren

1.72

auga

1.69

iphany

1.68

azo

1.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.