INDEX

Explanations

words associated with various forms of identity or labels

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

0:0.02

1:0.02

2:0.05

3:0.07

4:0.09

5:0.04

6:0.38

7:0.04

8:0.04

9:0.06

10:0.08

11:0.04

Negative Logits

KL

-1.44

VICE

-1.39

unknown

-1.27

NCT

-1.20

Jaw

-1.19

[*

-1.16

Kis

-1.15

 unsus

-1.15

Gaw

-1.13

atform

-1.11

POSITIVE LOGITS

vironment

1.50

vre

1.46

rification

1.40

ignt

1.31

peat

1.30

odox

1.26

ailability

1.24

uilding

1.24

 Cipher

1.19

atile

1.19

Activations Density 0.006%

words associated with various forms of identity or labels

No Comments

No Known Activations