INDEX

Explanations

words related to urban planning, development projects, and geographical locations

oai_token-act-pair · gpt-3.5-turbo

Triggered by @bot

character names and specific character traits

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-att-kk

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

0:0.01

1:0.04

2:0.06

3:0.24

4:0.02

5:0.02

6:0.14

7:0.08

8:0.05

9:0.12

10:0.07

11:0.11

Negative Logits

 redistributed

-1.15

finder

-1.06

<+

-1.03

imates

-1.02

 lobb

-1.01

elected

-1.00

orno

-0.99

imil

-0.97

 slic

-0.96

eton

-0.95

POSITIVE LOGITS

adesh

1.30

GMT

1.24

 Pradesh

1.21

oola

1.20

�

1.19

culosis

1.16

api

1.15

�

1.11

 Abbey

1.05

 Ferry

1.04

Activations Density 0.003%

words related to urban planning, development projects, and geographical locations

character names and specific character traits

No Known Activations