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

Head Attr Weights

0:0.09

1:0.06

2:0.07

3:0.09

4:0.08

5:0.08

6:0.07

7:0.07

8:0.10

9:0.09

10:0.07

11:0.07

Negative Logits

ée

-1.66

enery

-1.54

thumbnails

-1.53

Untitled

-1.51

 photograp

-1.43

idth

-1.42

 guests

-1.42

 photographed

-1.41

SPONSORED

-1.38

nexus

-1.38

POSITIVE LOGITS

acon

1.72

rans

1.71

agar

1.60

 wiser

1.57

oning

1.55

ween

1.53

minster

1.53

omics

1.52

ostic

1.51

AME

1.50

Activations Density 0.000%

No Known Activations

This feature has no known activations.