INDEX

Explanations

references to medical evaluations and treatments, particularly related to head injuries and concussions

New Auto-Interp

Top Features by Cosine Similarity

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L10_Hcat_z_lr1.20e-03_l11.30e+00_ds24576_bs4096_dc1.00e-05_rsanthropic_rie25000_nr4_v9.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

float32

Hook Name

blocks.10.attn.hook_z

Hook Layer

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Activation Function

relu

Embeds

IFrame

Link

Head Attr Weights

0:0.24

1:0.05

2:0.05

3:0.07

4:0.10

5:0.03

6:0.10

7:0.10

8:0.02

9:0.04

10:0.08

11:0.08

Negative Logits

ellar

-3.28

 Parish

-3.01

 Exodus

-2.96

 Paso

-2.92

 Luna

-2.91

 Luffy

-2.90

 Zoro

-2.83

overty

-2.83

 Sparrow

-2.82

 Evangel

-2.74

POSITIVE LOGITS

 concussion

7.61

cussion

4.86

 helmet

4.59

 neurolog

4.41

 Conc

4.30

 helmets

4.29

NHL

4.04

ussions

4.02

 Athlet

3.96

 hockey

3.94

Activations Density 0.010%

references to medical evaluations and treatments, particularly related to head injuries and concussions

No Known Activations