INDEX

Explanations

references to social groups and relationships

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

1:0.02

2:0.11

3:0.05

4:0.35

5:0.03

6:0.05

7:0.11

8:0.03

9:0.04

10:0.07

11:0.07

Negative Logits

 sunset

-1.52

skirts

-1.50

 disclaim

-1.47

emption

-1.43

 condoms

-1.43

 ceremonial

-1.40

 revers

-1.39

estones

-1.38

 passes

-1.35

 timeout

-1.33

POSITIVE LOGITS

wealth

1.74

 onlook

1.73

 purposes

1.72

 alike

1.70

 sensibilities

1.69

 newcomers

1.61

inar

1.60

hematically

1.53

ankind

1.52

OPLE

1.52

Activations Density 0.053%

references to social groups and relationships

No Comments

No Known Activations