INDEX

Explanations

instances of the word "more" and related terms signaling additional content or information

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

0:0.21

1:0.21

2:0.02

3:0.04

4:0.02

5:0.09

6:0.06

7:0.01

8:0.10

9:0.06

10:0.06

11:0.05

Negative Logits

 resil

-1.79

NetMessage

-1.79

otrop

-1.75

 Tuls

-1.74

illus

-1.64

raint

-1.61

oğ

-1.60

annabin

-1.59

 Slater

-1.59

 Shap

-1.58

POSITIVE LOGITS

►

1.81

¢

1.75

RELATED

1.68

 VIEW

1.68

Days

1.63

 EVENT

1.58

++)

1.56

icle

1.53

 sidx

1.53

Â

1.52

Activations Density 0.002%

instances of the word "more" and related terms signaling additional content or information

No Known Activations