INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L1_Hcat_z_lr1.20e-03_l18.00e-01_ds24576_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v5.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

float32

Hook Name

blocks.1.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

1:0.06

2:0.07

3:0.07

4:0.08

5:0.11

6:0.09

7:0.08

8:0.07

9:0.07

10:0.08

11:0.09

Negative Logits

ault

-1.95

aults

-1.83

objects

-1.74

 Objects

-1.73

obyl

-1.63

 Float

-1.63

ocene

-1.63

Interstitial

-1.62

 slides

-1.62

 Interface

-1.60

POSITIVE LOGITS

 political

1.72

icia

1.69

 HuffPost

1.68

hig

1.68

 Princ

1.67

 wedd

1.64

sylv

1.63

 princip

1.62

 regional

1.60

 sworn

1.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.