INDEX

Explanations

mentions of media outlets

New Auto-Interp

Top Features by Cosine Similarity

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L3_Hcat_z_lr1.20e-03_l19.00e-01_ds24576_bs4096_dc1.00e-06_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.3.attn.hook_z

Hook Layer

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Activation Function

relu

Embeds

IFrame

Link

Head Attr Weights

0:0.02

1:0.04

2:0.10

3:0.22

4:0.02

5:0.01

6:0.10

7:0.12

8:0.05

9:0.09

10:0.06

11:0.10

Negative Logits

ahime

-1.41

jriwal

-1.17

ザ

-1.14

pillar

-1.13

raped

-1.11

iour

-1.09

abase

-1.07

 Ashes

-1.04

 resemble

-1.03

utterstock

-1.03

POSITIVE LOGITS

backer

1.34

Balt

1.17

Outside

1.11

monary

1.11

noxious

1.09

 Admission

1.08

Elsewhere

1.08

 Brass

1.07

NB

1.06

ocaust

1.06

Activations Density 0.008%

mentions of media outlets

No Known Activations