INDEX

Explanations

emotional states and experiences related to hope and loss

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_L5_Hcat_z_lr1.20e-03_l11.00e+00_ds49152_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v9.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

Data Type

float32

Hook Name

blocks.5.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

1:0.01

2:0.06

3:0.05

4:0.12

5:0.03

6:0.08

7:0.43

8:0.03

9:0.03

10:0.05

11:0.05

Negative Logits

ansom

-1.52

isSpecialOrderable

-1.46

aughters

-1.39

uscript

-1.36

amia

-1.36

Lear

-1.35

 Tutorial

-1.34

adra

-1.33

 tutorials

-1.32

 answ

-1.31

POSITIVE LOGITS

 vibration

1.96

 vibrations

1.92

 discomfort

1.75

 contraction

1.67

 pinch

1.63

 recoil

1.55

aldehyde

1.54

 uneasy

1.50

 nausea

1.50

vib

1.47

Activations Density 0.062%

emotional states and experiences related to hope and loss

No Comments

No Known Activations