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

1:0.07

2:0.09

3:0.08

4:0.08

5:0.07

6:0.08

7:0.08

8:0.09

9:0.07

10:0.07

11:0.09

Negative Logits

icky

-1.71

 nonsense

-1.66

 intolerable

-1.61

 blight

-1.60

 Sinn

-1.59

 hallmark

-1.55

harm

-1.55

 senseless

-1.54

 trem

-1.48

 biod

-1.47

POSITIVE LOGITS

 Shroud

1.91

ysis

1.86

 undergone

1.71

gee

1.67

successfully

1.65

 Username

1.64

 Proper

1.56

 Received

1.54

esses

1.54

 Strength

1.53

Activations Density 0.000%

No Known Activations

This feature has no known activations.