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

1:0.06

2:0.09

3:0.08

4:0.07

5:0.08

6:0.07

7:0.09

8:0.09

9:0.08

10:0.09

11:0.08

Negative Logits

 ambul

-1.79

TextColor

-1.68

 Alban

-1.64

 Ukrain

-1.63

ogly

-1.61

 Polic

-1.55

uristic

-1.54

inav

-1.51

 antioxid

-1.51

 relig

-1.51

POSITIVE LOGITS

 racked

1.68

Jeremy

1.52

 Geoff

1.52

Buzz

1.51

 prolific

1.47

gd

1.43

advertisement

1.41

CHR

1.40

duration

1.40

Released

1.40

Activations Density 0.000%

No Known Activations

This feature has no known activations.