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

2:0.08

3:0.09

4:0.09

5:0.09

6:0.08

7:0.05

8:0.08

9:0.09

10:0.07

11:0.07

Negative Logits

bolt

-1.78

ノ

-1.78

antha

-1.72

dinand

-1.70

ornings

-1.69

EStream

-1.67

ixt

-1.66

UGH

-1.66

bsp

-1.64

方

-1.64

POSITIVE LOGITS

 unaccount

1.90

 legisl

1.79

 sanctioned

1.72

 monarchy

1.71

 outlawed

1.67

 graft

1.63

 legislatures

1.61

 surg

1.60

 overhaul

1.58

 loophole

1.57

Activations Density 0.000%

No Known Activations

This feature has no known activations.