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

1:0.06

2:0.10

3:0.05

4:0.06

5:0.04

6:0.06

7:0.25

8:0.07

9:0.05

10:0.12

11:0.06

Negative Logits

unknown

-1.60

 Recession

-1.58

 precip

-1.50

 Tomorrow

-1.48

 unseen

-1.48

 Evolution

-1.47

nown

-1.47

 Crossref

-1.46

 Recover

-1.45

 mirac

-1.45

POSITIVE LOGITS

ovie

1.78

mson

1.61

 decoration

1.61

estyles

1.57

 guideline

1.54

hemy

1.53

 stamp

1.48

amily

1.47

tex

1.46

ciating

1.46

Activations Density 0.000%

No Known Activations

This feature has no known activations.