INDEX

Explanations

references to digital communication services and social media platforms

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_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

Not in Any Lists

No Comments

Head Attr Weights

0:0.06

1:0.02

2:0.04

3:0.09

4:0.13

5:0.14

6:0.06

7:0.05

8:0.04

9:0.13

10:0.09

11:0.09

Negative Logits

¯¯

-2.62

 Ridley

-2.54

 Referred

-2.45

 Puppet

-2.34

龍

-2.28

 dwarves

-2.20

 Rober

-2.17

Rated

-2.13

okingly

-2.12

 guiActiveUnfocused

-2.10

POSITIVE LOGITS

 WhatsApp

4.60

 newsletter

3.26

 Telegram

3.21

Chat

2.92

 Whats

2.87

legram

2.82

ews

2.73

 newsletters

2.65

 Subscribe

2.65

scribe

2.63

Activations Density 0.002%

references to digital communication services and social media platforms

No Comments

No Known Activations