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

1:0.08

2:0.07

3:0.07

4:0.08

5:0.09

6:0.08

7:0.09

8:0.08

9:0.08

10:0.07

11:0.08

Negative Logits

 Stadium

-1.76

 capacitor

-1.48

abyte

-1.47

Pik

-1.44

illions

-1.43

 Shards

-1.42

 marrow

-1.40

 nanop

-1.40

 sink

-1.39

 SetTextColor

-1.37

POSITIVE LOGITS

OE

1.63

BUS

1.56

sole

1.52

 interchangeable

1.52

��

1.51

 guiActiveUn

1.50

VII

1.50

 veter

1.50

WARE

1.48

IRD

1.47

Activations Density 0.000%

No Known Activations

This feature has no known activations.