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

2:0.09

3:0.09

4:0.09

5:0.07

6:0.08

7:0.09

8:0.08

9:0.06

10:0.08

11:0.07

Negative Logits

assadors

-2.17

*/(

-2.06

adesh

-1.89

quart

-1.85

psey

-1.85

anooga

-1.78

ineries

-1.76

fman

-1.76

ospels

-1.75

signed

-1.74

POSITIVE LOGITS

 Garc

1.59

 Gupta

1.52

 Hogan

1.47

Sph

1.47

 silk

1.46

 Sloven

1.46

 Icar

1.45

 Bris

1.44

 Rahman

1.44

 Ambro

1.44

Activations Density 0.000%

No Known Activations

This feature has no known activations.