INDEX

Explanations

mentions of people's names or entities in a news or report-like context

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

patterns of historical or factual events involving severity or consequences

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

Not in Any Lists

No Comments

Head Attr Weights

0:0.02

1:0.07

2:0.06

3:0.02

4:0.02

5:0.05

6:0.07

7:0.06

8:0.04

9:0.05

10:0.05

11:0.44

Negative Logits

ー

-1.45

ptions

-1.38

ュ

-1.35

etc

-1.35

ens

-1.31

 Juda

-1.29

 rehears

-1.23

ements

-1.21

BUS

-1.21

apons

-1.19

POSITIVE LOGITS

 namely

1.44

nesday

1.31

opoly

1.24

gey

1.24

onite

1.24

:{

1.20

hire

1.17

ixty

1.13

icent

1.10

 worth

1.10

Activations Density 0.353%

mentions of people's names or entities in a news or report-like context

patterns of historical or factual events involving severity or consequences

No Comments

No Known Activations

mentions of people's names or entities in a news or report-like context

patterns of historical or factual events involving severity or consequences

No Comments

No Known Activations