INDEX

Explanations

words related to specific character names or identifiers

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

1:0.03

2:0.06

3:0.07

4:0.05

5:0.04

6:0.38

7:0.05

8:0.05

9:0.07

10:0.08

11:0.05

Negative Logits

 ILCS

-1.39

 unfocusedRange

-1.25

EMS

-1.17

STD

-1.16

 objective

-1.14

 Illum

-1.14

 threats

-1.14

-+-+

-1.13

itored

-1.12

)].

-1.09

POSITIVE LOGITS

doi

1.72

nova

1.48

mite

1.32

hoff

1.30

cloth

1.30

merce

1.28

hs

1.23

ν

1.22

lite

1.21

Stack

1.20

Activations Density 0.001%

words related to specific character names or identifiers

No Comments

No Known Activations