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

1:0.09

2:0.08

3:0.08

4:0.08

5:0.07

6:0.07

7:0.09

8:0.08

9:0.07

10:0.08

11:0.09

Negative Logits

agues

-2.04

 proble

-1.85

[/

-1.77

NetMessage

-1.76

ottesville

-1.75

erest

-1.75

-1.73

 livest

-1.72

 SHOULD

-1.69

)</

-1.68

POSITIVE LOGITS

 Held

1.57

 recount

1.54

 hammered

1.49

 Blanc

1.49

Hew

1.48

marked

1.48

 Chevron

1.45

 outraged

1.44

 frontman

1.43

 halted

1.42

Activations Density 0.000%

No Known Activations

This feature has no known activations.