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

0:0.07

1:0.12

2:0.07

3:0.08

4:0.09

5:0.06

6:0.07

7:0.09

8:0.07

9:0.07

10:0.08

11:0.09

Negative Logits

 Prop

-2.11

NV

-1.95

 Euros

-1.59

 reciproc

-1.58

VT

-1.54

 Persian

-1.53

 alternatively

-1.48

 Rouge

-1.45

 Gulf

-1.44

 differing

-1.44

POSITIVE LOGITS

ancies

1.87

ctors

1.85

mble

1.81

ancy

1.77

thood

1.76

jong

1.75

accompanied

1.71

avery

1.67

aturday

1.66

ocese

1.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.