INDEX

Explanations

phrases emphasizing proof or the need to demonstrate something

oai_token-act-pair · gpt-4o-mini Triggered by @bot

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

Not in Any Lists

No Comments

Head Attr Weights

0:0.01

1:0.03

2:0.13

3:0.15

4:0.01

5:0.03

6:0.04

7:0.18

8:0.07

9:0.12

10:0.07

11:0.10

Negative Logits

�

-1.25

cheon

-1.19

erv

-1.14

�

-1.08

Sut

-1.06

 Schwar

-1.03

Pg

-1.03

р

-1.02

�

-1.02

erest

-1.01

POSITIVE LOGITS

 validity

1.13

ihilation

1.12

rozen

1.12

orthy

1.11

oyal

1.11

querque

1.09

 rigged

1.06

 delinqu

1.05

 integrity

1.04

llah

1.03

Activations Density 0.011%

phrases emphasizing proof or the need to demonstrate something

No Comments

No Known Activations