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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.09

1:0.09

2:0.08

3:0.08

4:0.07

5:0.08

6:0.07

7:0.05

8:0.06

9:0.11

10:0.08

11:0.08

Negative Logits

Pac

-1.40

amping

-1.31

 Block

-1.30

Wonder

-1.30

pun

-1.30

cool

-1.26

block

-1.26

 Kimmel

-1.26

steel

-1.26

Chip

-1.21

POSITIVE LOGITS

otype

1.78

rift

1.68

awei

1.61

iasis

1.61

omsky

1.59

interstitial

1.53

 incorpor

1.51

 plaintiff

1.48

prus

1.47

cture

1.47

Activations Density 0.000%

No Known Activations

This feature has no known activations.