INDEX

Explanations

references to images or representations of individuals in various contexts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.03

1:0.02

2:0.07

3:0.08

4:0.34

5:0.04

6:0.06

7:0.09

8:0.05

9:0.05

10:0.06

11:0.07

Negative Logits

 Williamson

-1.50

 believes

-1.37

 understands

-1.32

quez

-1.30

 recognized

-1.29

 interpreted

-1.28

 suppose

-1.27

opal

-1.27

 emerged

-1.27

owder

-1.26

POSITIVE LOGITS

selves

1.49

Tact

1.47

 Goodbye

1.46

 goodbye

1.45

 Alone

1.42

xon

1.41

bye

1.39

xit

1.39

thood

1.39

anqu

1.38

Activations Density 0.000%

references to images or representations of individuals in various contexts

No Comments

No Known Activations