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

4:0.08

5:0.10

6:0.07

7:0.09

8:0.09

9:0.09

10:0.07

11:0.08

Negative Logits

opal

-1.86

 Defeat

-1.74

Inside

-1.72

 Thick

-1.71

throw

-1.70

Flash

-1.70

 Launcher

-1.66

Shadow

-1.63

rity

-1.61

heat

-1.59

POSITIVE LOGITS

 apprentices

1.92

 precept

1.81

 vocational

1.80

 voluntary

1.72

 duties

1.67

 salary

1.65

 teaching

1.62

GDP

1.58

 volunt

1.58

 monop

1.57

Activations Density 0.000%

No Known Activations

This feature has no known activations.