INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L2_Hcat_z_lr1.20e-03_l11.00e+00_ds24576_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v4.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

float32

Hook Name

blocks.2.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

1:0.07

2:0.07

3:0.08

4:0.07

5:0.08

6:0.08

7:0.09

8:0.08

9:0.09

10:0.07

11:0.08

Negative Logits

 Decay

-1.47

Story

-1.46

 Events

-1.44

 Vinyl

-1.42

 Classics

-1.42

 Receiver

-1.37

 Event

-1.37

 Ranked

-1.34

 Policies

-1.31

 Races

-1.31

POSITIVE LOGITS

gur

1.55

nap

1.48

heny

1.34

zin

1.32

hai

1.32

phrine

1.32

ipop

1.32

apan

1.31

roc

1.31

intel

1.31

Activations Density 0.000%

No Known Activations

This feature has no known activations.