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

IFrame

Link

Head Attr Weights

0:0.07

1:0.09

2:0.07

3:0.08

4:0.07

5:0.07

6:0.07

7:0.08

8:0.08

9:0.08

10:0.08

11:0.09

Negative Logits

 specialize

-1.64

 miniature

-1.59

 timer

-1.50

 slider

-1.44

 themed

-1.42

 glide

-1.41

 freezer

-1.41

 cleaned

-1.36

 transitioned

-1.36

 marker

-1.35

POSITIVE LOGITS

mining

1.86

 Juda

1.60

etheless

1.58

alion

1.57

覚醒

1.52

���

1.51

 Christians

1.50

 perjury

1.49

ribes

1.48

onym

1.47

Activations Density 0.000%

No Known Activations

This feature has no known activations.