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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.08

1:0.05

2:0.09

3:0.09

4:0.08

5:0.08

6:0.08

7:0.08

8:0.07

9:0.08

10:0.09

11:0.08

Negative Logits

 Syrians

-1.72

 Ukrainians

-1.72

 Britons

-1.59

 Filip

-1.55

VK

-1.54

 Russians

-1.53

Mug

-1.52

 inhab

-1.50

 Iranians

-1.50

 Bahrain

-1.50

POSITIVE LOGITS

ォ

1.86

使

1.79

ウス

1.78

══

1.77

ナ

1.74

etheless

1.72

ools

1.70

��

1.67

NEY

1.66

ERA

1.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.