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

0:0.04

1:0.04

2:0.05

3:0.07

4:0.06

5:0.04

6:0.18

7:0.26

8:0.04

9:0.04

10:0.08

11:0.06

Negative Logits

hement

-1.58

 defense

-1.58

semble

-1.57

 backbone

-1.53

essen

-1.49

ascus

-1.41

uled

-1.41

 vigil

-1.41

 memorial

-1.41

nod

-1.38

POSITIVE LOGITS

 GOODMAN

1.73

MpServer

1.69

ドラ

1.64

��

1.60

 inconsistency

1.54

 inacc

1.54

 Clicker

1.53

ALE

1.53

LOS

1.52

RANT

1.52

Activations Density 0.000%

No Known Activations

This feature has no known activations.