INDEX

Explanations

phrases related to subscriptions or updates from a news outlet

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

0:0.01

1:0.01

2:0.04

3:0.05

4:0.17

5:0.41

6:0.02

7:0.03

8:0.08

9:0.09

10:0.02

11:0.02

Negative Logits

 confir

-3.43

bc

-3.33

bec

-3.26

](

-3.17

*/(

-3.14

 favour

-2.95

[/

-2.85

ado

-2.72

yet

-2.66

?",

-2.61

POSITIVE LOGITS

pires

2.77

ensable

2.75

heon

2.72

phrine

2.68

oat

2.57

 Mayo

2.50

イト

2.50

if

2.48

etermin

2.36

 Chaser

2.35

Activations Density 0.007%

phrases related to subscriptions or updates from a news outlet

No Comments

No Known Activations