INDEX

Explanations

mentions of statistics and evidence related to gender representation issues

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

0:0.02

1:0.03

2:0.06

3:0.14

4:0.33

5:0.02

6:0.06

7:0.09

8:0.05

9:0.04

10:0.05

11:0.06

Negative Logits

iform

-1.67

skinned

-1.67

"},{"

-1.62

Cash

-1.53

ococ

-1.52

auri

-1.51

WARD

-1.46

must

-1.45

thens

-1.44

jong

-1.43

POSITIVE LOGITS

 limitations

1.98

 absence

1.95

 prohibition

1.89

 lack

1.82

 fact

1.76

 omission

1.76

 warnings

1.75

 negligence

1.74

 inaction

1.69

 limitation

1.68

Activations Density 0.019%

mentions of statistics and evidence related to gender representation issues

No Known Activations